Harvard Case Study: Central Bank Digital Currency

CSP045 J A N UA RY 2019 Cryptocurrency MA R G A RE T E. TA HY A R T RE V OR I. KI VI A T MA DI SO N J . R O BE RT S SUI W E N L I AN G H O W EL L J A CK SO N Memorandum TO: Deputy Director of Federal Reserve Bank Operations and Payment Systems at the Federal Reserve FROM: Director of Federal Reserve Bank Operations and Payment Systems DATE: February 2019 RE: Fed-Issued Digital Currency Introduction In light of the declining usage of cash within the United States, the Board of Governors of the Federal Reserve System requested that a team of staff economists brief the board on potential designs for a Federal Reserve digital fiat currency. The economist teams have suggested two possible designs— “Fedcoin” and “Fedcount.” There is an internal debate within the staff economists about which design might be better for payment efficiencies and monetary policy implementation, with some also expressing the view that no action should be taken by the Federal Reserve System at all. Both proposals are still in their early stages, as described further below. However, given that other central banks around the world have begun experimenting or considering experimentation with some central bank digital currencies (CBDCs), we have been asked to evaluate the three proposals: Fedcoin, Fedcount, and no action. To that end, I would like your help with analyzing the potential legal and policy issues that Written by Margaret E. Tahyar, Trevor I. Kiviat, Madison J. Roberts, and Suiwen Liang of Davis, Polk and Wardwell with assistance by Howell Jackson. Cases are developed solely as the basis for class discussion. They are not intended to serve as endorsements, sources of primary data, legal advice, or illustrations of effective or ineffective management. Copyright © 2019 President and Fellows of Harvard University. No part of this publication may be reproduced, stored in a retrieval system, used in a spreadsheet, or transmitted in any form or by any means—electronic, mechanical, photocopying, recording, or otherwise—without permission. To order copies or permissions to reproduce materials visit our website at casestudies.law.harvard.edu or by mail at Harvard Law School Case Studies Program, 1545 Massachusetts Avenue – Harvard Law School Library, 3rd Floor, Cambridge, MA 02138, or by email at HLSCaseStudies@law.harvard.edu. CRYPTOCURRENCY CSP045 could arise from the policy, monetary and legal issues that could arise either from the two proposals or from a decision that the Federal Reserve stays out of the fray. Your task is to consider any design recommendations to guide the Federal Reserve Board forward, carefully weighed against the merits and demerits of maintaining the status quo. Another work stream is tasked with analyzing whether the Federal Reserve System possesses legal authority for implementing CBDC in the absence of new legislation, so for now assume that such authority exists. You should, however, consider any policy issues related to the institutional independence of the Federal Reserve System and the views that the congressional oversight committee might have about decisions made by the Board of Governors with respect to CBDC and the role of the Federal Reserve System in the political economy. Background Functions of Money Sound money must fulfill three classical functions: medium of exchange, store of value, and unit of account. First, in order for an instrument to function as a medium of exchange, the instrument must be able to facilitate the sale of goods and services. The seller in a transaction must accept the instrument as a means of payment with the belief that the seller can give the instrument as a means of payment for other transactions. 1 Second, the instrument must serve as a store of value, preserving purchasing power over time. An instrument that is susceptible to depreciating or failing to maintain its value would not be considered sound money. 2 Third, functioning as a unit of account requires the instrument to act as a yardstick for measuring and comparing value across goods and services, thus informing the economic decisions of its users. 3 The Role of the Central Bank in the U.S. System The Federal Reserve System forms the central bank of the United States. The Federal Reserve System features “(1) a central governing Board, (2) a decentralized operating structure of 12 Federal Reserve Banks, and (3) a combination of public and private characteristics” 4 (for more information about the Federal Reserve System, see Appendix Item 3). The Board of Governors of the Federal Reserve System is based in the nation’s capital and supervises the 12 Federal Reserve Banks. The 12 Federal Reserve Banks “service[] financial institutions in 12 Federal Reserve districts.” 5 The Federal Reserve notes, what is commonly called cash, are issued by the separate regional Federal Reserve Banks but printed by the U.S. 1 BANK FOR INT’L SETTLEMENTS, ANNUAL ECONOMIC R EPORT, June 2018, at 92 (2018), https://www.bis.org/publ/arpdf/ar2018e.pdf [https://perma.cc/H4ER-SAGF]. (Hereinafter BIS ANNUAL ECONOMIC REPORT.) 2 Id. at 91. 3 Id. 4 About the Federal Reserve System, Structure of the Federal Reserve System, https://www.federalreserve.gov/aboutthefed/structurefederal-reserve-system.htm [https://perma.cc/V7ZF-RZM2]. 5 MICHAEL S. BARR, HOWELL E. JACKSON & MARGARET E. TAHYAR, FINANCIAL REGULATION: LAW AND POLICY 939 (2d ed. 2018). 2 CRYPTOCURRENCY CSP045 Treasury. The 12 Federal Reserve Banks also “act as banker’s banks, providing a wide variety of services such as storing currency and processing checks and electronic payments for both banking institutions and the federal government.” 6 Directly accessing these payment services “requires a master account at the nearest regional Federal Reserve Bank.”7 The Federal Reserve Banks only grant master accounts to banks. The Federal Reserve System performs a number of critical roles for the U.S. economy. First, the Federal Reserve Board is charged by Congress to “pursu[e] the goals of ‘maximum employment, stable prices, and moderate long-term interest rates.’” 8 Second, “[t]he Federal Reserve is the primary organ responsible for carrying out U.S. monetary policy and, for many, that is its most crucial role.” 9 Third, the Federal Reserve’s circulation of cash, issued and delivered by the regional Federal Reserve Banks to meet domestic and foreign demand, also generates significant revenues for the U.S. government—thanks to the U.S. dollar’s status as the global reserve currency. Forms of Money Conventional Forms of Money in the U.S. Financial System In the current U.S. financial system, money conventionally manifests in three forms: cash, commercial bank deposits, and central bank accounts—accounts held by banks at each of the Federal Reserve Banks through what is called a master account. As described below, the relationship between these three kinds of money is overseen by the Federal Reserve System. Source: Adapted from Bank for International Settlements10 6 Id. 7 Id. at 183. 8 Id. at 49. 9 Id. at 939. 10 Bank for Int’l Settlements, Central Bank Digital Currencies 5 (2018), https://www.bis.org/cpmi/publ/d174.pdf [https://perma.cc/7FQEHYLE]. (Hereinafter CENTRAL BANK DIGITAL CURRENCIES.) 3 CRYPTOCURRENCY CSP045 Cash consists of both metal coins minted and paper notes printed by the government. Cash represents a liability of the central bank 11 and is “legal tender for all debts, public charges, taxes, and dues.” 12 The Federal Reserve Banks issue and supply cash to commercial banks in their districts, which is then circulated by commercial banks to users based on demand. Generally, anyone can access, store, and use cash. Due to its widely recognized authenticating features (for example, watermarks and holograms on higher value bills), payments involving cash typically do not require a trusted third party to record the transfer or verify the authenticity of the physical notes or coins. 13 Consequently, settling transactions with cash is usually immediate and generally does not come with transaction fees. 14 Cash also comes with costs attached for both the user, who bears the burden of storage and loss, and the government, which bears the burden of supporting the printing, minting, and initial delivery infrastructure. Although cash is convenient for settling smaller transactions immediately, it is inconvenient for large transactions. Holding and transferring large quantities of cash is both burdensome and unsafe. For these reasons, the proportion of cash in payment values has been declining in many major economies, including the United States. 15 From the government’s perspective, cash requires an expensive infrastructure to support its circulation and upkeep. 16 In addition to the logistics of producing, storing, and transporting physical cash to accommodate new demand, the government has to periodically retire and replace unfit coins and notes to maintain existing supply, further consuming resources. 17 Even with these expenses, however, the government’s creation of cash generates significant profits for the United States government, due in part to the U.S. dollar’s role as a reserve currency and the widespread use of $100 bills as a safe haven asset or as the currency most valued in illicit transactions. Money in bank deposit accounts consists of “electronically recorded deposit account liabilities on the ledgers of commercial banks.”18 This form of money is universally available to anyone with a bank account. 19 Money stored in deposit accounts serves as “the main means of payment between ultimate users” 20 and “the main form of money holding of households and businesses.” 21 Its supply increases when 11 WALTER ENGERT & BEN S.C. FUNG, BANK OF CAN., CENTRAL BANK DIGITAL CURRENCY: MOTIVATIONS AND IMPLICATIONS 1 (2017). 12 31 U.S.C. § 5103. 13 Id. at 1. 14 BARR, JACKSON & TAHYAR, supra note 5, at 808. 15 Id. 16 Aleksander Berentsen & Fabian Schär, The Case for Central Bank Electronic Money and the Non-case for Central Bank Cryptocurrencies, 100 FED. RES. BANK OF ST. LOUIS REV. 97, 101 (2018). 17 For reference, there is an average life expectancy of 5.9 years for $1 bills, 4.2 years for $10, 3.7 years for $50, and 15 years for $100. How Currency Gets into Circulation, FED. RES. BANK OF N.Y. (July 2013), https://www.newyorkfed.org/aboutthefed/fedpoint/fed01.html [https://perma.cc/8CLH-WHYT]. 18 Ole Bjerg, Designing New Money – The Policy Trilemma of Central Bank Digital Currency 15 (Copenhagen Bus. Sch. Working Paper, June 14, 2017), https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2985381 [https://perma.cc/W7SH-L99G]. 19 Id. at 14–15. As a practical matter, however, “[m]ore than 9 million U.S. households, including 15.6 million adults and 16.3 million children are unbanked, or lack an account at an insured depository institution.” BARR, JACKSON & TAHYAR, supra note 5, at 826. There are “[a]nother 24.5 million households, comprising 51.1 million adults and 16.3 million children [who] are underbanked, meaning that although they have bank accounts, they also obtain financial services from non-bank, alternative-financial-services providers such as check cashers or payday lenders. Being unbanked or underbanked presents significant challenges for participating in many payment systems.” Id. 20 BIS ANNUAL ECONOMIC REPORT, supra note 1, at 93. 21 Bjerg, supra note 18, at 12. 4 CRYPTOCURRENCY CSP045 commercial banks issue loans to borrowers or receive cash deposits and decreases when account holders “make debt repayments or interest payments to the bank.” 22 Money in central bank accounts is recorded digitally as liabilities on each Federal Reserve Bank’s ledger.23 As discussed earlier, direct access to central bank money is currently restricted to chartered banks and other depository institutions. Nonbank companies and individuals cannot directly access the accounts at a Federal Reserve Bank for payments or storage. 24 Supply of central bank money grows when commercial banks purchase government bonds or deposit funds with their Federal Reserve Banks and contracts when commercial banks borrow on a collateralized basis from their Federal Reserve Banks. 25 Instead of facilitating retail payments, the amounts deposited by commercial banks with the Federal Reserve Banks enable the settlement of wholesale interbank payments. 26 Taken together, these three conventional forms of money share overlapping features while retaining key distinctions. For instance: • Cash and commercial bank deposits are, in principle, accessible to anyone, unlike master accounts, 27 which can only be opened by banks. • Cash and master accounts are liabilities issued by the Federal Reserve Banks, unlike commercial bank deposits. 28 • Commercial bank deposits and master accounts exist in digital form, unlike cash. 29 Beyond these abstract comparisons, the three forms of money share an operational infrastructure through the Federal Reserve System’s partnership with the banking sector in circulating cash to the public. 30 Specifically, individuals can receive cash by converting digital money stored in their bank deposits into paper notes, most typically by a withdrawal from an ATM machine. These paper notes are purchased by the bank from its regional Federal Reserve Bank through a deduction from the amount held in the bank’s master account and a corresponding delivery of physical notes. 31 22 Id. at 15. 23 See id. at 14. 24 If the Fintech Charters were granted master accounts, then access would be extended beyond depository institutions. See FINTECH CHARTER CASE STUDY 6–7. 25 See Bjerg, supra note 18, at 15. 26 BIS ANNUAL ECONOMIC REPORT, supra note 1, at 93. 27 Bjerg, supra note 18, at 15–16. 28 Id. 29 Id. 30 This institutional arrangement is not confined to the Federal Reserve System. “[I]n almost all modern-day economies, money is provided through a joint public-private venture between the central bank and private banks, with the central bank at the system’s core.” BIS ANNUAL ECONOMIC REPORT, supra note 1, at 93. 31 JP KONING, FEDCOIN: A CENTRAL BANK-ISSUED CRYPTOCURRENCY 25 (2016). 5 CRYPTOCURRENCY CSP045 New Forms of Money? Source: Adapted from Bank for International Settlements32 Cryptocurrency Cryptocurrencies, such as Bitcoin, Ethereum, and Ripple, are a type of digital currency that rely on cryptography to verify and secure transactions, as well as to manage the creation of new units.33 In contrast to the conventional forms of money, and in contrast to other digital currencies (for example, ingame currencies used by certain video game franchises), cryptocurrencies generally are not backed by a trusted institution and typically are not liabilities of any person or institution. 34 Distributed ledger technology, of which blockchain technology forms a subset, represents the technological engine underpinning many cryptocurrencies, including Bitcoin. Distributed ledger technology “refers to the protocols and supporting infrastructure that allow computers in different locations to propose and validate transactions and update records in a synchronized way across a network.” 35 As the term ledger implies, distributed ledger technology provides a means of recording account balances or transaction history. In most other contexts, electronic transactions are recorded on a centralized ledger. Generally, a trusted intermediary (e.g., the central bank, commercial banks, or PayPal) manages the central ledger to “track account holders’ balances and, ultimately, vouch for a 32 CENTRAL BANK DIGITAL CURRENCIES, supra note 10, at 5. 33 BARR, JACKSON & TAHYAR, supra note 5, at 835. For information on Bitcoin in particular, see Reuben Grinberg, Bitcoin: An Innovative Alternative Digital Currency, 4 HASTINGS SCI. & TECH. L.J. 159, 160 (2011). 34 Gov. Lael Brainard, Bd. of Governors of the Fed. Reserve Sys., Cryptocurrencies, Digital Currencies, and Distributed Ledger Technologies: What Are We Learning? (May 15, 2018), https://www.federalreserve.gov/newsevents/speech/brainard20180515a.htm [https://perma.cc/NY4Z-SJNV]. 35 Morten Linnemann Bech & Rodney Garratt, Central Bank Cryptocurrencies, BIS Q. REV., Sept. 2017, at 55, 58, https://www.bis.org/publ/qtrpdf/r_qt1709f.pdf [https://perma.cc/4YH3-DXXV]. 6 CRYPTOCURRENCY CSP045 transaction’s authenticity.” 36 Not so with distributed ledger technology. As the name further implies, the ledger is distributed across computers and other internet-connected devices in separate locations globally, without the need for a trusted central authority. 37 For Bitcoin and many other cryptocurrencies, “[t]his ledger is the blockchain.” 38 Transactions are recorded in batches, or “blocks,” with new blocks being “chained” in order to amend the existing ledger with additional transactions. 39 This process of clearing and settlement happens around the clock—24/7/365— and it all occurs mathematically, with minimal human intervention. Parties wishing to transact with blockchain technology must announce their transaction “to the entire network, effectively asking network participants to determine its authenticity.” 40 Responsibility for verifying the validity of new blocks is shared by nodes—or computers connected to the network—through a consensus mechanism, whereby the nodes agree to the common state of ledger usually with cryptographic tools and protocol rules. 41 For Bitcoin’s proof-of-work consensus mechanism, network participants compete to solve cryptographic puzzles necessary for validating a new block. As an economic incentive, the first to succeed receives newly issued units of bitcoin. 42 A critical distinction to keep in mind is that Bitcoin is an example of a permissionless system, where each node possesses a complete and current copy of the ledger. In principle, anyone can participate in validating transactions in a permissionless system. In contrast, a permissioned system only allows trusted nodes—in other words, participants approved by a central entity—to participate in updating the ledger. A permissioned system may involve additional access controls such as verification of identity. 43 Many harbor doubts as to whether permissionless cryptocurrencies can function as sound money—much less supplant the use of cash. For instance, Professor David Yermack argued in 2013 that Bitcoin, which remains the most popular cryptocurrency, failed to satisfy the classical criteria of money. Although Bitcoin enjoys some acceptance as a form of payment, “the worldwide commercial use of bitcoin remains miniscule. . . .”44 He maintained that Bitcoin performs poorly as a unit of account since Bitcoin-based quotes for prices of ordinary goods commonly extend to “four or five decimal places with leading zeros, a practice rarely seen in consumer marketing and likely to confuse both sellers and buyers in the 36 Trevor I. Kiviat, Beyond Bitcoin: Issues in Regulating Blockchain Transactions, 65 DUKE L.J. 569, 578 (2015); BIS ANNUAL ECONOMIC REPORT, supra note 1, at 96. 37 As such, this system is a “trustless” system. 38 Id.; see also SATOSHI NAKOMOTO, BITCOIN: A PEER-TO -PEER ELECTRONIC CASH SYSTEM 3 (2009), https://bitcoin.org/bitcoin.pdf [https://perma.cc/8B5G-5RGK]. 39 BANK FOR INT’L SETTLEMENTS, DISTRIBUTED LEDGER TECHNOLOGY IN PAYMENT, CLEARING AND SETTLEMENT 3 (2017), https://www.bis.org/cpmi/publ/d157.pdf [https://perma.cc/FXS4-A2E7]. (Hereinafter DISTRIBUTED LEDGER TECHNOLOGY IN PAYMENT, CLEARING AND SETTLEMENT). 40 Kiviat, Beyond Bitcoin, supra note 36, at 578. 41 DISTRIBUTED LEDGER TECHNOLOGY IN PAYMENT, CLEARING AND SETTLEMENT, supra note 39, at 3–4. 42 Id. at 4. Bitcoin is programmed to have a finite total outstanding supply. When Bitcoin creation ceases, “the incentive to validate transactions will likely be transaction fees.” Kiviat, Beyond Bitcoin, supra note 36, at 579–580. 43 BIS ANNUAL ECONOMIC R EPORT, supra note 1, at 96. 44 David Yermack, Is Bitcoin a Real Currency? An Economic Appraisal 2 (Nat’l Bureau of Econ. Research, Working Paper No. 19747, 2013). 7 CRYPTOCURRENCY CSP045 marketplace.”45 Even then, his most severe reservation resided with Bitcoin’s prospects as a store of value given, among other factors, its high volatility. 46 Since 2013, use of Bitcoin and other cryptocurrencies as a medium of exchange has increased, especially for those without access to or wishing to bypass central banks and commercial banks. Yet, adoption remains far from widespread. 47 More problematically, persisting volatility continues to bely cryptocurrency’s function as a stable means for storing value. To many observers, cryptocurrency’s extreme fluctuations and growing number harken back ominously to the era of wildcat banking, when state banks circulated their own currency that too often had dubious worth (see Appendix Item 11). 48 Central Bank Digital Currency The advent of permissionless cryptocurrencies has catalyzed guarded interest among academics and policymakers in CBDC. In a recent survey conducted by the Bank for International Settlements (see Appendix Item 7), 70% of the responding central bank participants reported current or imminent engagement with CBDC work. 49 Although a number of central banks are researching CBDC,50 few have announced plans to implement CBDC in the next decade. 51 Despite the growing body of literature, CBDC currently is “not a well-defined term”52 and standardization has yet to be reached in the form of an agreed taxonomy or lexicon. Since discussions on CBDC have been mostly conceptual, it might be most helpful to define CBDC by way of contrast: CBDC is central bank-issued digital money that is distinct from the existing master accounts at Federal Reserve Banks. 53 Conventional forms of money New forms of money Bank Master account Privately issued Central bank Cash deposits balances cryptocurrency digital currency Digital × ✔ ✔ ✔ ✔ Central bank-issued × × ✔ ✔ ✔ Universally accessible × (✔) ✔ ✔ ✔ ✔ = existing or likely feature, (✔) = possible feature, × = not typical or possible feature Source: Adapted from Bank for International Settlements54 45 Id. at 2–3. That said, there have been proposals to introduce a millibitcoin (mBTC), then worth $1.85, to better account for pricing of conventional goods, Kai Sedgwick, It’s Time to Change the Way We Measure Bitcoin, BITCOIN.COM (Dec. 8, 2017), https://news.bitcoin.com/its-time-to-change-the-way-we-measure-bitcoin/ [https://perma.cc/XB67-K7AY]. 46 As discussed in footnote 42, Bitcoin’s supply is ultimately fixed. Although in theory the consensus mechanism could vote to lift this limit, doing so would be difficult in practice. This effective limit on supply presents a further challenge to Bitcoin’s viability as a form of money. 47 CHRISTIAN BARONTINI & HENRY HOLDEN, BANK FOR INT’L SETTLEMENTS, PROCEEDING WITH CAUTION – A SURVEY ON CENTRAL BANK DIGITAL CURRENCY 14 (2019), https://www.bis.org/publ/bppdf/bispap101.pdf [https://perma.cc/B3B2-H35Z]. 48 Robert C. Hockett, Money’s Past Is Fintech’s Future: Wildcat Crypto, the Digital Dollar, and Citizen Central Banking (Dec. 11, 2018), https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3299555 [https://perma.cc/7XBF-838X]. 49 BARONTINI & HOLDEN, supra note 47, at 7. 50 Id. at 12. 51 Id. at 7. 52 CENTRAL BANK DIGITAL CURRENCIES, supra note 10, at 3. 53 Id. at 4. 54 Id. at 6. 8 CRYPTOCURRENCY CSP045 CBDC is also distinct from privately issued cryptocurrencies: CBDC would be backed by the government in the same way that current forms of fiat currency are and may assuage concerns of volatility because supply could be set programmatically (or algorithmically). 55 Some commentators regard a hypothetical CBDC as superior to permissionless cryptocurrencies in two respects. First, CBDC would be a more reliable and stable store of value. 56 Second, CBDC would likely avoid the “significant waste of resources” required by the consensus mechanisms used by some cryptocurrencies. 57 For these reasons, a few commentators advocate CBDC as a solution for counteracting the migration of users from conventional forms of money to cryptocurrencies. 58 Nevertheless, no central bank has yet reported curbing the propagation of cryptocurrency as an urgent priority. 59 More broadly, the measured pace of concrete action suggests that central bank interest in CBDC remains tempered by caution. 60 Our own central bank has expressed its reservations more openly, stating in May 2018 that “there is no compelling demonstrated need for a Fedissued digital currency” 61 (see Appendix Item 1). Design Choices for Central Bank Digital Currency Whether to introduce a CBDC and its optimal design features depend on the objectives and motivations of the central bank. 62 Designing a CBDC would need to take into consideration the following features, among others. Currencies can either be token-based or account-based. If token-based, careful thought should be given to the appropriate degree of anonymity. As with all currencies, CBDC would also require an infrastructure to support its distribution (centralized or decentralized). As a digital currency, CBDC must also have in place a validation scheme (centralized or decentralized) to prevent double spending or identity theft. Finally, digital currencies can be subject to caps and/or accrue interest. 55 Brainard, supra note 34. 56 Id. 57 DONG HE ET AL., IMF, VIRTUAL C URRENCIES AND BEYOND: INITIAL CONSIDERATIONS, IMF STAFF DISCUSSION NOTE 22 n.20 (Jan. 2016) (“Mining Bitcoin is costly, requiring computer processing power and associated energy costs. In addition, such systems involve a negative externality that causes overinvestment in computer power.”). 58 Hockett, supra note 48. See also Itai Agur, Central Bank Digital Currencies: An Overview of Pros and Cons, in DO WE NEED CENTRAL BANK DIGITAL CURRENCY? 113, 115 (2018) (“[O]ne incentive that central banks may have to develop a retail CBDC is to limit demand for private cryptocurrencies.”). 59 BARONTINI & HOLDEN, supra note 47, at 14. 60 Id. at 12–13. 61 Brainard, supra note 34. 62 NOTE: While the Bank of Canada has undertaken research and experimentation with CBDCs for interbank settlement (wholesale CBDC), our interests concern retail payments. To that end, please focus on the features that may be conducive towards a retail or general-purpose CBDC, which may differ from those for a wholesale CBDC. 9 CRYPTOCURRENCY CSP045 Conventional forms of money New forms of money TokenAccountBank Master account Privately issued Cash based based deposits balances cryptocurrency CBDC CBDC Tokens or accounts T A A T T A Anonymity × × × ✔ ✔ (✔) Decentralized distribution × × ✔ ✔ ✔ (✔) Decentralized validation × (✔) (✔) ✔ ✔ ✔ Capped × × × × (✔) (✔) Interest-bearing × ✔ ✔ (✔) (✔) (✔) ✔ = existing or likely feature, (✔) = possible feature, × = not typical or possible feature Source: Adapted from Bank for International Settlements63 Technology: Tokens or Accounts The technological vehicle for the CBDC could be token-based, involving the transfer of an object of value from one wallet into another, or account-based, involving the transfer of a claim recorded on one account to another. 64 Cash and Bitcoin are examples of token-based money, whereas bank accounts and master accounts are examples of account-based money. One distinction between tokens and accounts lies in the method of verifying an exchange: the focus of verification for token-based money is the object transferred—i.e., the token—and the focus of verification for account-based money is the identities of the account holders. 65 Transfers involving tokens depend on the payee’s ability to “verify the validity of the payment object,” 66 whether in the form of a physical or digital coin. Token-based systems must control for counterfeiting and enable the payee to validate the authenticity of the received token. 67 This implication holds true for both physical and digital coins, as the payer could use a fake token in a transaction. Digital currencies introduce the additional problem of double spending, where the payer uses a real digital token for more than one transaction. Thus, policymakers for token-based systems must grapple with designing validation channels (e.g., via a recognizable design as with paper notes or via a decentralized consensus mechanism as with Bitcoin) for limiting counterfeit tokens and duplicate tokens. In contrast, transfers involving accounts rely on verifying “the identity of the account holder.” 68 Account-based systems must control for identity theft and the unauthorized transfer or withdrawal of money held within valid accounts. 69 Consequently, policymakers for account-based systems must seek ways to validate the identity of the transacting parties. 70 63 CENTRAL BANK DIGITAL CURRENCIES, supra note 10, at 6. 64 TOMMASO MANCINI-GRIFFOLI ET AL., IMF, CASTING LIGHT ON CENTRAL BANK DIGITAL CURRENCY, IMF STAFF DISCUSSION NOTE 7 (Nov. 2018). 65 CENTRAL BANK DIGITAL CURRENCIES, supra note 10, at 4. 66 Id. 67 Id. 68 Id. 69 Id. 70 Id. (“Identification is needed to correctly link payers and payees and to ascertain their respective account histories.”). 10 CRYPTOCURRENCY CSP045 Anonymity As seen with Bitcoin, token-based systems rely on verifying the authenticity of the exchanged token, not the identities of the transacting parties. The payer “need reveal nothing to the payee beyond the information associated with the specific coin.” 71 As such, a token-based CBDC can “be designed to provide different degrees of anonymity”72 for its users or traceability for its transactions. 73 That said, Bitcoin and Ethereum would be more accurately characterized as granting pseudo-anonymity. While “[t]he blockchain does not record real names or physical addresses,” the transactions of the ledger are public and would be traceable to the owner should “the owner of the wallet become[] known.” 74 A paper trail may be harder to follow than a digital one. 75 Cash transactions are usually anonymous to third parties (such as banks) and the government. 76 Indeed, the anonymity of cash is an attractive quality for many as a medium for protecting the privacy of their transaction histories. The degree to which a digital token should be—or even could be—designed to be anonymous with respect to (i) the counterparty, (ii) third-party validators, and (iii) the government remains an arena of lively debate. In contrast, account-based systems generally require some knowledge of the transacting parties’ identities, such as the unique account number of the other party. 77 However, even if there was relative counterparty anonymity—for example, where the parties only knew the other’s account numbers—thirdparty anonymity is likely absent. As discussed below, the banks operating the accounts would be “required to have information regarding the individuals’ identities for a variety of legal reasons.”78 Degree of Centralization for Distribution and Validation The central bank could opt to (i) directly oversee and manage the CBDC or (ii) delegate roles to other actors, such as commercial banks. For a token-based CBDC, the Federal Reserve Banks could directly handle the distribution of new digital tokens to consumers and/or directly operate the validation process of digital tokens as a central validating node. Alternatively, the Federal Reserve Banks could collaborate with private entities to accomplish these responsibilities. For distribution, the Federal Reserve Banks could partner with commercial banks to circulate the digital tokens to their consumers based on demand. This decentralized distribution scheme is already used for the circulation of cash. As for validation, the Federal Reserve Banks could rely on a network of nodes outside the Federal Reserve System, be it a permissionless network or a permissioned network, to prevent double spending and preserve the integrity of the ledger. 71 Charles M. Kahn, Francisco Rivadeneyra & Tsz-Nga Wong, Should the Central Bank Issue E-Money? 11 (Oct. 2018), https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3271654 [https://perma.cc/6KST-XJVT]. 72 CENTRAL BANK DIGITAL CURRENCIES, supra note 10, at 4. 73 Mancini-Griffoli et al., supra note 64, at 4. 74 Ankit Panda, Cryptocurrencies and National Security, COUNCIL ON FOREIGN RELS. (Feb. 28, 2018), https://www.cfr.org/backgrounder/cryptocurrencies-and-national-security. 75 See Agur, supra note 58, at 115 (“An essential feature of physical cash is its anonymity.”). 76 This is less often the case with respect to the transacting counterparty, since the cash transactions typically take place in person. 77 Kahn, Rivadenyra & Wong, supra note 71, at 11. 78 Id. 11 CRYPTOCURRENCY CSP045 Although an account-based CBDC would be held by Federal Reserve Banks, the degree to which the Federal Reserve Banks directly administer those accounts would remain a deliberate choice. A more centralized scheme would involve the Federal Reserve Banks designing and operating the account’s verification requirements and associated payment and customer services. Alternatively, the Federal Reserve Banks could delegate the day-to-day responsibilities of administering the digital accounts to private sector firms, such as commercial banks. 79 Quantitative Limits The Federal Reserve Board and the Federal Reserve Banks could impose quantitative limits on transaction or storage sizes for CBDC “as a way of controlling potentially undesirable implications or to steer usage in a certain direction.” 80 For instance, a cap could be imposed on the amount of CBDC that can be stored in a wallet or account. Alternatively, there could be a cap in the amount of CBDC a user can transact in for a single transaction. Finally, softer limits could be imposed, where the user would be permitted to hold or transact in an amount of CBDC beyond the limit—but with reduced anonymity. Interest-bearing Unlike cash, digital tokens and accounts could be designed to pay interest (positive or negative.)81 Indeed, existing forms of account-based money, such as master accounts and commercial bank accounts, are already interest-bearing. Importing this characteristic to digital tokens is also technically feasible. Positive CBDC interest rates would encourage storing CBDC and converting holdings of other currencies into CBDC; negative CBDC interest rates would encourage spending and converting holdings of CBDC into other currencies. 82 Legal and Policy Considerations The particular bundle of features that define a CBDC will pose important legal and policy implications for the central bank. Provided below are a few to consider. Concerns Related to Anti-Money Laundering, Know Your Customer, and Counter-Terrorism Financing The Federal Reserve Board would need to consider concerns and policies relating to laws on anti-money laundering (AML) and counter-terrorism financing (CFT). Banks are subject to various laws that restrict them from providing financial services that would assist with criminal activity. Such laws also require banks to maintain customer due diligence programs for bank accounts and monitor suspicious activity, 79 See BEN DYSON & GRAHAM HODGSON, POSITIVE MONEY, DIGITAL CASH: WHY CENTRAL BANKS SHOULD START ISSUING ELECTRONIC MONEY 16–19 (2016). 80 CENTRAL BANK DIGITAL CURRENCIES, supra note 10, at 4. 81 A negative interest rate means that the user pays the central bank to store its CBDC. Negative interest rates are highly controversial and viewed by many as a confiscation of private property. Since a direct deduction would be made from the account, negative interest rates are quite different from the erosion of inflation. 82 CENTRAL BANK DIGITAL CURRENCIES, supra note 10, at 6. 12 CRYPTOCURRENCY CSP045 including cash transaction amounts exceeding $10,000. 83 AML and know-your-customer laws (KYC) also apply to the Federal Reserve Banks, but current compliance is easy in practice, as only banks have direct access to master accounts with Federal Reserve Banks. CBDC that grants broad anonymity to users and limits traceability of transactions could become a favored medium for illicit activities, thereby potentially raising legal and reputational concerns for the Federal Reserve System. Cryptocurrencies have been criticized by some as a preferred medium of payment for facilitating illegal activities. 84 A recent Europol report found that cryptocurrencies “remain the primary payment mechanism for the payment of criminal services. . . .”85 For example, the Islamic State of Iraq and al-Sham (ISIS), a militant terrorist organization, solicited donations in Zcash and Bitcoin in order to finance their propaganda efforts. 86 Cryptocurrencies also figure prominently in so-called Darknet markets, which are online marketplaces for illicit commodities and services, especially drugs. 87 A recent paper concluded that “illegal activity accounts for a substantial proportion of the users and trading activity of bitcoin.” 88 Such concerns caution against issuing opaque digital tokens, at least not without a key in the government’s hands to unlock encrypted transaction information. On the other hand, there may be legitimate demands for anonymity to safeguard privacy. Transacting parties might seek to avoid the unwelcome nuisance of directed advertising or weightier dangers of identity theft or personal harm. 89 Purchase history could also reveal health-related conditions and shopping habits, which, while not unlawful, could cause embarrassment. After all, “knowledge by a third party of the payee, amount, and time of payment for every transaction made by an individual can reveal a great deal about the individual’s whereabouts, associations and lifestyle.” 90 The trends towards privacy around consumer data means that users are increasingly worried about what happens to their information after its collection, since data could be shared, sold, or lost. Indeed, the anonymity of cash remains an attractive quality for many as a medium for protecting the privacy of their transaction histories. Even if 83 31 C.F.R. § 1010.311 (“Each financial institution . . . shall file a report of each deposit, withdrawal, exchange of currency or other payment or transfer, by, through, or to such financial institution which involves a transaction in currency of more than $10,000. . . .”). 84 Large-denominated paper notes—which grant almost complete anonymity—are also widely used for transactions related to criminal activities, including abroad, see Kenneth S. Rogoff, Response to Jeffrey Rogers Hummel’s Review of The Curse of Cash, 14 ECON. J. WATCH, May 2017, at 164, 168 (“[W]hile there are many reasonable uses of the $100 bill abroad, it is indisputably popular with Russian oligarchs, Mexican drug lords, illegal arms dealers, Latin American rebels, corrupt officials, human traffickers, etc., and of course North Korean counterfeiters.”). 85 EUROPOL, INTERNET ORGANISED CRIME THREAT ASSESSMENT (IOCTA) 2018, at 58 (2018). 86 Id. at 53 (“Cryptocurrencies represent a source of opportunity for terrorist groups, allowing them to move funds across borders while avoiding the regular banking scrutiny. . . . [By] the end of 2017 . . . [Islamic State] sympathisers triggered mass cryptocurrency (Bitcoin and the more anonymous Zcash) donation campaigns in [Islamic State] affiliated websites as well as in chat environments (e.g. Telegram) to support their cause.”). 87 Id. at 47. In 2014, the FBI and DEA shut down Silk Road, perhaps the most infamous Darknet market. Id. 88 Sean Foley, Jonathan R. Karlsen & Tālis J. Putniņš, Sex, Drugs, and Bitcoin: How Much Illegal Activity Is Financed Through Cryptocurrencies?, OXFORD L.: OXFORD BUS. L. BLOG (Feb. 19, 2018), https://www.law.ox.ac.uk/business-law-blog/blog/2018/02/sex-drugs-and-bitcoin-howmuch-illegal-activity-financed-through [https://perma.cc/G9B6-BMPB] (“For example, approximately one-quarter of all users (25%) and close to one-half of bitcoin transactions (44%) are associated with illegal activity. The estimated 24 million bitcoin market participants that use bitcoin primarily for illegal purposes (as at April 2017) annually conduct around 36 million transactions, with a value of around $72 billion, and collectively hold around $8 billion worth of bitcoin. To give these numbers some context, the total market for illegal drugs in the US and Europe is estimated to be around $100 billion and €24 billion annually.”). 89 Bech & Garratt, supra note 35, at 64. 90 DAVID CHAUM, Blind Signatures for Untraceable Payments, in ADVANCES IN CRYPTOLOGY, PROCEEDINGS OF CRYPTO 82 199, 199 (1983). 13 CRYPTOCURRENCY CSP045 the government or company collecting the data fully aligned with the user’s interests, others—such as foreign states or rogue hackers—could also be trying to get ahold of that information. Monetary Policy Provocatively, CBDC could expand the Federal Reserve’s arsenal for controlling monetary policy in two respects. For the purposes of this briefing, please consider how CBDC’s implications for monetary policy relate to discussions on whether consumers’ usage of currency should fall outside or within the control of the central bank. 91 This analysis should include how the controversial nature of these policies may impact the political economy. First, replacing cash with an interest-bearing digital currency could grant the Federal Reserve a powerful new instrument for effectuating negative interest rates 92—which means that the user pays the central bank to store its currency—a policy theorized to discourage users from hoarding money and stimulate spending in the economy. Because cash has no interest rate, a central bank’s ability to cut interest rates becomes constrained when it approaches negative interest rate territory, as people can convert their holdings in the banking system into cash. However, if cash were to be supplanted by an interest-bearing CBDC, the Federal Reserve could overcome this effective lower bound. 93 Second, account-based CBDCs would enable the Federal Reserve to proactively airdrop funds into the accounts of consumers as another way of encouraging spending. 94 Some argue that such “helicopter drops” would operate more directly and precisely than reliance on quantitative easing, interest rate adjustments or commercial bank lending. 95 Furthermore, helicopter drops can be performed universally on all accounts or on a selective basis, depending on the relevant situation. 96 These options would come at severe political cost, which you should account for in your briefing preparations. First, the imposition of negative interest rates assumes a feasible path towards eliminating cash. As a matter of practical reality, the abolition of cash and the threat of negative interest rates would confirm the suspicions held by broad demographic swaths of the country about the Federal Reserve’s ambitions for repressing financial autonomy (see Appendix Item 13 and Appendix Item 18). Second, helicopter drops overtly tread into controversial decisions about redistribution—for example, decisions 91 NOTE: An economic policy division is handling the precise implications on monetary policy so no need for you to closely parse the economic technicalities or the merits of deploying the following policies for the purposes of our briefing. 92 See Kenneth Rogoff, Dealing with Monetary Paralysis at the Zero Bound, 31 J. OF ECON. PERSPS., Summer 2017, at 47, 57–58 (2017). 93 See id. 94 Some suggest that helicopter drops could also be performed by token-based CBDC, see, e.g., Mike Bird, HSBC Says the Blockchain Could Be Used for “Helicopter Money,” BUS. INSIDER (Nov. 9, 2015), https://www.businessinsider.com/hsbc-says-the-blockchain-could-be-used-forradical-central-bank-helicopter-money-policies-2015-11?r=UK&IR=T [https://perma.cc/7PT9-KJDZ], but this mechanism would likely be less straightforward, see DYSON AND HODGSON, supra note 79, at 22 (“[I]t would be extremely easy for the Bank of England to make small and regular ‘helicopter drops’ to every citizen, as a tool of monetary policies.”). 95 DYSON & HODGSON, supra note 79, at 2 & 8. 96 For example, the government could selectively use helicopter drops for lower-income households to stimulate spending and “cushion[] their purchasing power from the effects of the downturn as well as from the temporarily negative level of the CBDC interest rate,” Michael D. Bordo & Andrew T. Levin, Central Bank Digital Currency and the Future of Monetary Policy 3 n.10 (Hoover Institution, Econ. Working Paper 17104, Aug. 2017); see also DYSON & HODGSON, supra note 79, at 8; MANCINI-GRIFFOLI ET AL., supra note 64, at 16 n.22 (“[Helicopter drops] would not necessarily reach all citizens.”). 14 CRYPTOCURRENCY CSP045 about how much money should be delivered and to whom. 97 Both actions would provoke public backlash and reignite critiques of the democratic legitimacy and institutional independence of the Federal Reserve. 98 Seigniorage Seigniorage, the profit made by the central bank from its issuance of currency (a function of the currency’s face value minus production and distribution costs), 99 forms a vital source of revenue for the Federal Reserve System and the U.S. Treasury (that is the American taxpayer). 100 Because of its status as the leading international reserve currency, the U.S. dollar generates substantial seigniorage revenues compared to other national currencies. 101 Seigniorage is particularly pronounced for high-denomination bills (namely $100 bills), which account for nearly 80% of the total value of U.S. dollars in supply 102 and also enjoy high demand outside the United States. 103 The Federal Reserve System does not rely on congressional appropriations, 104 which could fluctuate based on political pressures. The independent funding of the Federal Reserve provides an additional safeguard for its institutional autonomy and independent policymaking. 105 Seigniorage revenues would decline if demand for U.S. currency decreases. Cash’s share in payment amounts has dropped considerably over recent years. Its decline could continue if users begin to favor cryptocurrencies or foreign currencies over U.S. currency, whether in the form of paper notes or digital notes. Although for now these concerns remain far from compelling, further developments in 97 MANCINI-GRIFFOLI ET AL., supra note 64, at 16 n.22 (“[T]he issue of legitimacy remains: how does the central bank decide how much to transfer to each household given the notable and very explicit redistributional consequences? Finally, helicopter drops would continue to be viewed as a form of monetary financing, thus undermining central bank independence.”). 98 See BARR, JACKSON & TAHYAR, supra note 5, at 49 (“The Federal Reserve Board’s power, however, remains controversial, and many critics want the Federal Reserve Board weakened or abolished. Critics allege that the Federal Reserve Board is secretive and undemocratic. . . .”); see also PAUL TUCKER, UNELECTED POWER: THE QUEST FOR LEGITIMACY IN CENTRAL B ANKING AND THE REGULATORY STATE (2018). 99 ENGERT & FUNG, supra note 11, at 3. 100 See Who Owns the Federal Reserve?, FAQs, FED. RES., https://www.federalreserve.gov/faqs/about_14986.htm [https://perma.cc/Q8GL- XHLR] (“[T]he Federal Reserve Banks are required by law to transfer net earnings to the U.S. Treasury, after providing for all necessary expenses of the Federal Reserve Banks, legally required dividend payments, and maintaining a limited balance in a surplus fund.”). 101 To illustrate, the Federal Reserve Banks transferred to the U.S. Treasury $97.7 billion, $91.5 billion, $80.2 billion, and $62.2 billion (estimated) in earnings remittances in 2015, 2016, 2017, and 2018, respectively. Board of Governors of the Federal Reserve System, Press Release: Federal Reserve Board announces Federal Reserve Bank income and expense data and transfers to the Treasury for 2018 (Jan. 10, 2019), https://www.federalreserve.gov/newsevents/pressreleases/other20190110a.htm [https://perma.cc/QPH5-8SUS]. 102 Kenneth S. Rogoff, The Curse of Cash 3 & 51 (2016). 103 See Ruth Judson, Big Note, Small Note: Central Bank Digital Currency and Cash, in DO WE NEED CENTRAL BANK DIGITAL CURRENCY? 33, 35 (2018) (“[T]he estimates in Judson (2016) indicate that about 70% of $100s could be held abroad, with the rest held at home.”); ROGOFF, THE CURSE OF CASH, supra note 102, at 39 (“[F]oreign holdings . . . might explain as much as 50% of US dollar holdings.”). 104 What Does It Mean that the Federal Reserve Is “Independent Within the Government”?, FAQs, FED. RES., https://www.federalreserve.gov/faqs/about_12799.htm [https://perma.cc/8FYQ-JVNE] (Mar. 1, 2017) (“The Federal Reserve does not receive funding through the congressional budgetary process.”); Who Owns the Federal Reserve?, FAQs, FED. RES., https://www.federalreserve.gov/faqs/about_14986.htm [https://perma.cc/3M4A-8MNM]. 105 Christopher J. Waller, Independence + Accountability: Why the Fed Is a Well-Designed Central Bank, FED. RES. BANK OF ST. LOUIS REV. 293, 298 (2011) (“[A] common method for politicians to entice government agencies is to threaten to cut the agencies’ budgets. . . . To counteract this possibility, Congress gave the Federal Reserve budget autonomy when it created the Fed in 1913. The Fed was given the power to earn its own income and spend it without government interference.”). 15 CRYPTOCURRENCY CSP045 stablecoins—currently mostly experimental 106—and non-U.S. CBDCs—currently mostly conceptual 107— could provide an appealing alternative to the U.S. dollar. On the other hand, some have argued that developing a U.S. CBDC could recover leakage of or even exceed the seigniorage from U.S. cash.108 This, of course, hinges on the perhaps overoptimistic assumption that future demand for a U.S. CBDC would match or exceed the existing demand for U.S. cash. Indeed, replacing cash with a CBDC that fails to appeal to users could inadvertently hasten the decline of seigniorage. Financial Stability Some academics worry that a retail CBDC that resembles bank deposits could raise the costs of deposittaking for commercial banks, thereby reducing their ability to perform productive lending. Since CBDC would be backed by the U.S. government, it would pose a safer alternative to bank deposits, which would be guaranteed only up to the deposit insurance limit of $250,000 per depositor. 109 There is a policy question whether it is sensible for the U.S. government to provide what would be, in effect, unlimited insurance of deposit funds. In times of crisis, the qualities of CBDC having the “safety of physical cash but convenience of bank deposit accounts” 110 could induce depositors to flee from depository institutions towards the central bank, thereby weakening financial stability.111 Indeed, the traditional bank run has been caused by depositors converting bank deposits into cash. A conversion of bank deposits into CBDC would have a similar impact on bank balance sheets and stability. Unlike a traditional bank run, the speed and scale would likely be unprecedented since a digital run could be triggered at the click of a button. 112 Others, however, are less worried about financial stability. Some, including Professor Morgan Ricks, embrace the perceived structural implications of expanding the central bank’s role and causing “largescale migration from bank deposits” to the central bank (see Appendix Item 14). 113 These proponents take the view that radical disintermediation of the private banking sector from deposit-taking would generate more financial stability, not less. The goals of these projects also include disintermediating the private banking sector from the lending side as well. 106 Stablecoins seek to overcome the volatility endemic among cryptocurrencies. Specifically, they are a type of cryptoasset for which the value is pegged to the value of another asset, such as the U.S. dollar, gold, or algorithmic pricing based on circulation supply. Santiago Fernández de Lis, Central Bank Digital Currencies: Features, Options, Pros and Cons, in DO WE NEED CENTRAL BANK DIGITAL CURRENCY? 46, 54 (2018). 107 BARONTINI & HOLDEN, supra note 47, at 1 (“The survey shows that, although a majority of central banks are researching CBDCs, this work is primarily conceptual and only a few intend to issue a CBDC in the short to medium term.”). 108 See DYSON & HODGSON, supra note 79, at 11–12; see also Morgan Ricks, John Crawford & Lev Menand, A Public Option for Bank Accounts (or Central Banking for All) 16–17 (Vanderbilt Univ. Law Sch., Research Paper No. 18-33, 2018), https://ssrn.com/abstract=3192162 [https://perma.cc/24MT-8U5Q]. 109 But see MANCINI-GRIFFOLI ET AL., supra note 64, at 25 (commenting that while deposit insurance does not immunize banks from runs, it significantly mitigates the risk). 110 DYSON & HODGSON, supra note 79, at 27. 111 CENTRAL BANK DIGITAL CURRENCIES, supra note 10, at 16. 112 Id. 113 Morgan Ricks, John Crawford & Lev Menand, A Public Option for Bank Accounts (or Central Banking for All) 2 & 17 (Vanderbilt Univ. L. Sch., Research Paper No. 18-33, 2018), https://ssrn.com/abstract=3192162 [https://perma.cc/7H4V-4J7Q]. While Professor Ricks and his coauthors would prefer “more traditional nomenclature,” id. at 25–26, his FedAccount proposal is an example of a universally accessible, accounts-based CBDC. 16 CRYPTOCURRENCY CSP045 Cybersecurity While offering convenience and efficiencies, digitization also carries risks of cyberthreats, including malware and fraud. 114 Cybersecurity is a significant operational risk and central banks are not immune to hacks. 115 In fact, instituting a CBDC could elevate the profile of the central bank as a target for cyberattacks. 116 The likelihood and severity of cyberattacks would be further exacerbated for a CBDC that enjoys reserve currency status and is universally accessible and would open the platform to “many participants and points of attack.” 117 Options for a U.S. Central Bank Digital Currency The two proposed CBDCs aim to ease the convenience and costs of payment systems and enhance monetary policy implementation. With this background in mind, please consider the three options discussed below. Both proposed CBDCs would be issued by Federal Reserve Banks, denominated in U.S. dollar, and convertible with other forms of money; similar to cash, both CBDCs would be deemed legal tender. As a third option, consider whether it would be more prudent to adopt a wait-and-see approach and stick with the existing system. Option 1: Fedcoin Fedcoin is a digital token built on a permissioned blockchain (“Fedchain”) 118 and issued by the Federal Reserve Banks. Unlike permissionless cryptocurrency, like Bitcoin, the production of Fedcoin would be managed and controlled by the Federal Reserve Banks, which would also serve as the only trusted parties in the network. Federal Reserve Banks would possess the ability to create and destroy Fedcoin in order to preserve a 1:1 conversion ratio between Fedcoin and the dollar. Although Federal Reserve Banks would maintain Fedcoin’s value, approved nodes would maintain the ledger’s verification, validating new transactions and screening out counterfeiting and double spending. These nodes would be operated by a select group of large commercial banks approved by the Federal Reserve Banks. Users could not directly access Fedcoin through an account at a Federal Reserve Bank, but instead must withdraw Fedcoin from their commercial bank accounts. Once withdrawn, the Fedcoin would become anonymous to the bank. Subsequently, Fedcoin could be stored in digital wallets provided by various private sector firms (e.g., banks or fintech companies) certified by a Federal Reserve Bank. Anyone who purchases and installs the requisite wallet software into a smartphone or personal computer would be able to store and pay with Fedcoin. A user could set up as many addresses for the wallet as he or she wishes. Once the wallet is set up, a user could engage in transactions in a similar way to Bitcoin. 114 CENTRAL BANK DIGITAL CURRENCIES, supra note 10, at 10. 115 Other Federal institutions have demonstrated vulnerability to cyberattacks. See, e.g., Jim Sciutto, OPM Government Data Breach Impacted 21.5 Million, CNN (July 10, 2015), https://www.cnn.com/2015/07/09/politics/office-of-personnel-management-data-breach-20million/index.html [https://perma.cc/XQK8-JZ99] (“Government investigators now believe that the data theft from the Office of Personnel Management computer systems compromised sensitive personal information, including Social Security numbers, of roughly 21.5 million people from both inside and outside the government, the government announced Thursday.”). 116 Brainard, supra note 34. 117 CENTRAL BANK DIGITAL CURRENCIES, supra note 10, at 10. 118 NOTE: Among other things, Fedchain would overcome Bitcoin’s technological constraints of scalability. 17 CRYPTOCURRENCY CSP045 As mentioned before, the nodes would be operated by large commercial banks approved by the Federal Reserve Banks and would validate Fedcoin transactions between users. These entities would also use Fedcoin for their own payments. As an incentive for operating these nodes, those responsible for validation will be able to collect transaction fees. If consensus is reached by these nodes, the approved Fedcoin transaction would be recorded on the Fedchain. Although Fedchain would be a public ledger, only transaction amounts and party addresses would be viewable. Option 2: Fedcount Fedcount offers a new account-based money created and held by the Federal Reserve Banks. Unlike master accounts, Fedcounts would be generally accessible to the nonbank public for holding electronic money. Fedcounts would hold electronic money for all users who register with their Federal Reserve Bank. To prevent fraud and enable instantaneous verification, registration with a Federal Reserve Bank and login to Fedcount would require fingerprint and/or facial recognition, which is already technologically feasible with smartphones. After successful login, users could review the account balance and transaction history of the Fedcount but nothing else. The development of expanded interfaces for payments and other functions would have to be provided by private sector firms, such as commercial banks. Such institutions would administer the relevant services to make Fedcount suitable for a user’s needs in exchange for fees. Integral functions to develop include the interfaces to initiate payments and to review more detailed transaction summaries. Other functions would include internet coverage, periodic statements, and customer support. Although a commercial bank might administer Fedcounts, Fedcounts would be distinct from a deposit account at a private bank because Fedcounts would be directly held at a Federal Reserve Bank and belong to the account holder (i.e., the user). The relevant Fedcount administrator would only be responsible for offering services associated with the Fedcount. Therefore, unlike with conventional bank accounts, commercial banks providing Fedcounts would not be able to conduct lending with money housed within Fedcounts. Moreover, money stored in Fedcount would represent a liability of the relevant Federal Reserve Bank, not of the Fedcount administrator. Thus, according to the analysts, even if the bank administering services to the Fedcount were to fail, the Fedcount would remain safe with the Federal Reserve Bank. Option 3: Neither For the purposes of practicably advancing payment efficiencies and monetary policy implementation, neither Fedcoin nor Fedcount would be superior to maintaining the existing infrastructure supporting cash. It is therefore prudent to continue with the current system for the foreseeable future. 18 CRYPTOCURRENCY CSP045 Briefing Questions Having examined the proposals, please review the materials included in the attached Appendix and prepare to brief the director. In particular, the director is eager to know your thoughts as to the following: 1. What are the comparative strengths and weaknesses of Fedcoin and Fedcount? In what ways can these proposals be improved? 2. What would be the likely reaction from financial sector stakeholders to these proposals? Consider the following four examples: • Ames Bank is a depository institution located in Cambridge, Massachusetts. Its business primarily consists of making loans to businesses, for which much of the funding is done through deposit-taking. • BitBank is a cryptocurrency company that supports and develops wallets for Bitcoin and Bitcoinderived currencies. Its wallets do not have compatibility with other types of cryptocurrencies. • The Zcash Company supports and develops Zcash, which is a privacy-protecting digital currency. Zcash features zero-knowledge proofs that allow the payee to prove the validity of a transaction without revealing information about the transaction itself. This allows transactions to be fully shielded from being traced within the public blockchain, thereby completely protecting the users’ privacy. • Circle is a payment technology service that provides wallet services. Recently, the company also purchased Poloniex, a prominent token exchange platform, as a step towards its goal of becoming a trading exchange for cryptocurrencies. 3. Consider the following proposals made by our IT specialists: • For Fedcoin – IT specialists claim that Fedchain’s cybersecurity capabilities could be enhanced by moving from a permissioned network of nodes to a permissionless network of nodes. They argue that a permissionless network would offer more robust operational resilience as the consensus mechanism could continue to operate should any node become unavailable or compromised. What are the benefits and concerns? • For Fedcount – IT specialists insist that Fedcount has the technological capability of providing more comprehensive services and greater financial inclusion for users if the Federal Reserve assumed control over administering Fedcount’s payment and customer services. What are the benefits and concerns? 4. Although cash is deemed legal tender, Federal law does not obligate a private business to accept cash payments. Businesses retain discretion to accept payment in whatever form they prefer. Take for example, bus lines that refuse pennies or convenience stores that refuse high-denominated bills.119 Other countries such as China and France take a different approach, making it generally unlawful for payees to refuse notes and coins with the status of legal tender. What are the benefits and drawbacks of these two approaches as applied to CBDC? How likely is the U.S. approach to change to this 119 Legal Tender Status, U.S. Dep’t of Treasury (Jan. 4, 2011), https://www.treasury.gov/resource-center/faqs/currency/pages/legal- tender.aspx [https://perma.cc/76QN-GBYN]. 19 CRYPTOCURRENCY CSP045 alternative approach? How should society weigh the payer’s right to choose between payment methods with that of the payee’s? 5. As the readings suggest, many closely associate anonymity with privacy. Yet, the two are distinct. Information can be private but not anonymous, such as transaction logs with personally identifying information safeguarded by a bank; likewise, information could be anonymous but not private, such as the sharing or disclosure of spending patterns aggregated across large numbers of users. Is this a meaningful distinction? If so, how does it pertain to the conversation about CBDC? 6. How does CBDC, whether by its nature or by design choice, compare with the attributes of cash? 7. Circulation and usage of U.S. notes extends beyond U.S. citizenship and territoriality. On the other hand, eligibility for master accounts with Federal Reserve Banks is limited to U.S. entities with a few exceptions. What legal and policy considerations would be implicated if CBDC access is granted internationally or restricted nationally? If the latter, should any controls be in place? 8. How would you rank the three options for purposes of payment efficiencies and monetary policy implementation? • Would your conclusion change if the motivation for developing CBDC was instead to advance (i) financial stability or (ii) financial inclusion (see footnote 19)? Is either financial stability or financial inclusion a pressing objective for the Federal Reserve System? 20 CRYPTOCURRENCY CSP045 Appendix Speech by Governors of the Federal Reserve Board • Item 1: Remarks by Gov. Lael Brainard, Board of Governors of the Federal Reserve System, Cryptocurrencies, Digital Currencies, and Distributed Ledger Technologies: What Are We Learning? (May 15, 2018), https://www.federalreserve.gov/newsevents/speech/brainard20180515a.htm [https://perma.cc/4C96-79UN]. Attached. Congressional Hearing – The Future of Money: Digital Currency, Hearing Before the Subcommittee on Monetary Policy and Trade of the Committee on Financial Services, US House of Representatives (July 18, 2018). • Item 2: Transcript Excerpts of the Prepared Statement of Eswar S. Prasad. Attached. Casebook – Michael S. Barr, Howell E. Jackson & Margaret E. Tahyar, Financial Regulation: Law and Policy (2d ed. 2018). • Item 3: Excerpts of Chapters 1.2 and 9.1 IMF Publications • Item 4: Christine Lagarde, Managing Director, IMF, Winds of Change: The Case for New Digital Currency (Nov. 14, 2018). Attached. • Item 5: Tommaso Mancini-Griffoli et al., IMF, Casting Light on Central Bank Digital Currency, IMF Staff Discussion Note (Nov. 2018). Attached. BIS Report • Item 6: Bank for International Settlements, Central Bank Digital Currencies (2018), https://www.bis.org/cpmi/publ/d174.pdf [https://perma.cc/E6ZB-4XHE]. Attached. • Item 7: Christian Barontini & Henry Holden, Bank for International Settlements, Proceeding with Caution – A Survey on Central Bank Digital Currency (2019), https://www.bis.org/publ/bppdf/bispap101.pdf [https://perma.cc/E3Y5-PZ8K]. Attached. Academic Articles • Item 8: Itai Agur, Central Bank Digital Currencies: An Overview of Pros and Cons, in Do We Need Central Bank Digital Currency? (2018). Not attached for reasons of copyright. • Item 9: Aleksander Berentsen & Fabian Schär, The Case for Central Bank Electronic Money and the Non-case for Central Bank Cryptocurrencies, 100, Federal Reserve Bank of St. Louis Rev. 97 (2018). Not attached for reasons of copyright. • Item 10: Charles M. Kahn, Francisco Rivadeneyra & Tsz-Nga Wong, Should the Central Bank Issue E-Money? (Oct. 2018), https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3271654 [https://perma.cc/N8YN-CHNZ]. Not attached for reasons of copyright. 21 CRYPTOCURRENCY CSP045 • Item 11: Robert C. Hockett, Money’s Past Is Fintech’s Future: Wildcat Crypto, the Digital Dollar, and Citizen Central Banking (Dec. 11, 2018), https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3299555 [https://perma.cc/72HWXEXM]. Not attached for reasons of copyright. • Item 12: Kenneth Rogoff, Dealing with Monetary Paralysis at the Zero Bound, 31 Journal of Economic Perspectives, Summer 2017, at 47 (2017). Not attached for reasons of copyright. • Item 13: Jeffrey Rogers Hummel, The War on Cash: A Review of Kenneth Rogoff’s The Curse of Cash, 14 Econ Journal Watch, May 2017, at 138. Not attached for reasons of copyright. • Item 14: Morgan Ricks, John Crawford & Lev Menand, A Public Option for Bank Accounts (or Central Banking for All) (Vanderbilt University Law School, Research Paper No. 18-33, 2018), https://ssrn.com/abstract=3192162 [https://perma.cc/74U5-KUBF]. Not attached for reasons of copyright. News Articles • Item 15: Izabella Kaminska, When the State Takes on the Digital Float, the State Takes on the Risk, Financial Times (Jan. 10, 2019), https://ftalphaville.ft.com/2019/01/10/1547096400000/When-the-state-takes-on-the-digitalfloat--the-state-takes-on-the-risk/. Not attached for reasons of copyright. • Item 16: Nouriel Roubini, Central Bank Digital Currencies Will Destroy Cryptocurrencies, ING Think (Nov. 19, 2018), https://www.project-syndicate.org/commentary/central-banks-takeover-digital-payments-no-cryptocurrencies-by-nouriel-roubini-2018-11 [https://perma.cc/DA7SS7WQ]. Not attached for reasons of copyright. • Item 17: Simon Scorer, Central Bank Digital Currency: DLT, or not DLT? That Is the Question, Bank Underground (June 5, 2017), https://bankunderground.co.uk/2017/06/05/central-bankdigital-currency-dlt-or-not-dlt-that-is-the-question/ [https://perma.cc/UGY2-3VMG]. Not attached for reasons of copyright. • Item 18: James Grant, Hostage to a Bull Market, Wall Street Journal (Sept. 9, 2016), https://www.wsj.com/articles/hostage-to-a-bull-market-1473456611. Not attached for reasons of copyright. 22 For release on delivery 7:00 p.m. EDT [4:00 p.m. PDT Local Time] May 15, 2018 Cryptocurrencies, Digital Currencies, and Distributed Ledger Technologies: What Are We Learning? Remarks by Lael Brainard Member Board of Governors of the Federal Reserve System at the Decoding Digital Currency Conference Sponsored by the Federal Reserve Bank of San Francisco San Francisco, California May 15, 2018 It is a pleasure to be here today. What better place to discuss digital currencies than in San Francisco, home to so many technology innovators working on new ways to disrupt various aspects of our daily lives? 1 Because of the transformative potential of digital currency and distributed ledger technologies, the Federal Reserve is actively monitoring digital innovations in the financial system. We have been keenly evaluating developments in fintech and digital currencies through a multidisciplinary lens, combining information technology and policy analysis to study their potential implications for payments policy, supervision and regulation, financial stability, monetary policy, and the provision of financial services. This work draws from expertise throughout the Federal Reserve System and benefits from engagement with our colleagues internationally. Cryptocurrencies The past decade has seen a wave of important new developments in digital technologies for payments, clearing, and settlement. Cryptocurrencies represent the leading edge of this digital wave. And it was the advent a decade ago of Bitcoin, the first cryptocurrency, that first gave shape to the vision of a decentralized digital currency. At the heart of any cryptocurrency is the creation of a new type of asset--the unit of the cryptocurrency itself--that is distinct from any traditional form of money used in routine transactions, such as U.S. currency or checking accounts in commercial banks. A typical cryptocurrency would not be a liability of any individual or institution. There is no trusted institution standing behind it. This is in stark contrast to U.S. currency and reserve balances, 1 I am grateful to David Mills of the Federal Reserve Board for his assistance in preparing this text. The remarks represent my own views, which do not necessarily represent those of the Federal Reserve Board or the Federal Open Market Committee. -2which are liabilities of the Federal Reserve Banks, and deposit accounts, which are liabilities of a bank or another regulated depository institution backed by federal insurance up to a specific level. And while a typical cryptocurrency may be used in payments, it is not legal tender, in contrast to U.S. currency. A typical cryptocurrency relies on the use of distributed ledger technology, which provides a new way to keep ownership records and transfer ownership from one user to another, often with little to no information about the identity of the owner. For instance, Bitcoin relies on the blockchain, which is run by anonymous computers all over the world linked together through a ledger of anonymized transactions. Digital currencies use automation via computer processing power, networking via the internet, and cryptography to transfer value from one person to another. What is innovative is that the computer code behind these transactions uses automated checks and balances to validate the sender and receiver, and whether there is enough value in the sender’s account to make the payment. Traditionally, this validation would be done by banks and payment networks. Instead, with a cryptocurrency, this validation could be done by anyone with enough computing power and resources to participate. Importantly, this technology is not owned or managed by any entity--regulated or not--that would be responsible for its maintenance, security, and reliability. Rather, its maintenance, security, and reliability are handled by a decentralized developer community, which often lacks strong governance. This combination of a new asset, which is not a liability of any individual or institution, and a new recordkeeping and transfer technology, which is not maintained by any single individual or institution, illustrates the powerful capabilities of today’s technologies. But there are also serious challenges. For instance, cryptocurrencies have exhibited periods of extreme volatility. If you purchased Bitcoin in December 2017 at a value of over $19,000, your -3electronic claims would be worth close to half that today. 2 Indeed, Bitcoin’s value has been known to fluctuate by one-quarter in one day alone. Such extreme fluctuations limit an asset’s ability to fulfill two of the classic functions of money: to act as a stable store of value that people can hold and use predictably in the future, and to serve as a meaningful unit of account that can be used to assign a comparable value of goods and services. In addition to losses, individual investors should be careful to understand the potential for other risks. 3 Cryptocurrencies may raise important investor and consumer protection issues. The lack of strong governance and questions about the applicable legal framework for some cryptocurrencies may make consumers vulnerable to mistakes, thefts, and security breaches without much, or any, recourse. Although the cryptographic technology may be robust to some events, such as the fraudulent double spending of the same units of the cryptocurrency for more than one transaction, the large number of breaches at some cryptocurrency exchanges and wallet providers suggest that significant vulnerabilities may remain with respect to security protections around customers’ accounts. 4 These breaches remind us that relying solely on cryptography within the transfer technology is not enough. Ultimately, a more holistic approach to the security of the broader cryptocurrency ecosystem, along with added layers of security on top of cryptography, are likely to be necessary for cryptocurrencies to be widely adopted. Some cryptocurrencies also appear quite vulnerable to money-laundering (BSA/AML, or Bank Secrecy Act/anti-money-laundering) concerns. Since many cryptocurrencies store in their 2 See, for example, https://www.coinbase.com/charts. Lael Brainard, “An Update on the Federal Reserve’s Financial Stability Agenda” (speech delivered at the Center for Global Economy and Business, Stern School of Business, New York University, New York, NY, April 3, 2018), https://www.federalreserve.gov/newsevents/speech/brainard20180403a.htm. 4 For example, Coincheck, a Tokyo-based cryptocurrency exchange was hacked in 2018. See https://www.wsj.com/articles/cryptocurrency-worth-530-million-missing-from-japanese-exchange-1516988190. A similar attack occurred back in 2014 to another Tokyo-based cryptocurrency exchange, Mt. Gox. See https://www.wsj.com/articles/mt-gox-to-hold-news-conference-1393579356. 3 -4ledger little to no information about the identity of owners of the cryptocurrency, this essentially mimics a bearer instrument--that is, an instrument whereby the holder of the instrument is presumed to be its owner. Further, cryptocurrencies are easy to transfer across borders. Indeed, a cryptocurrency that mimics a bearer instrument and provides significant anonymity in transactions, including across borders, could raise significant concerns regarding the potential to facilitate illicit activities and associated money laundering. For example, electronic instruments can be easily transferred and stored in large amounts, and peer-to-peer transactions outside of the United States could be very hard to prevent and detect. Such instruments appear to have proven susceptible for use to convey payments to illicit actors--for example, to pay ransoms. Overall, however, the still relatively small scale of cryptocurrencies in relation to our broader financial system and relatively limited connections to our banking sector suggest that they do not currently pose a threat to financial stability. 5 Of course, if cryptocurrencies were to achieve wide-scale use, or their impact were greatly magnified through leverage, the effects could be broader. In particular, adverse developments and shifts in sentiment could cause a global rush to exit this market. As we have seen in other speculative activity in the past, rushfor-the-exits behavior can aggravate price fluctuations, create trading difficulties, and even induce market breakdowns. Thus, we will continue to monitor cryptocurrencies as they evolve, with particular vigilance for any signs of growing materiality to the broader financial system. Central Bank Digital Currencies Given some of the inherent issues and challenges that cryptocurrencies pose for investor and consumer protection and the prevention of money laundering, some have advocated that central banks should create their own digital forms of currency as more stable and reliable 5 See https://g20.org/sites/default/files/media/communique_-_fmcbg_march_2018.pdf. -5alternatives to cryptocurrencies. After all, a central bank digital currency could overcome the volatility risks associated with an unbacked asset with no intrinsic value by substituting a digital instrument that is the direct liability of the central bank. Moreover, advocates suggest a central bank would be able to develop a transfer mechanism that has robust governance. Even though central bank digital currencies may at first glance appear to address a number of challenges associated with the current crop of cryptocurrencies, this appeal may not withstand closer scrutiny. 6 First, there are serious technical and operational challenges that would need to be overcome, such as the risk of creating a global target for cyberattacks or a ready means of money laundering. For starters, with regard to money laundering risks, unless there is the technological capability for effective identity authentication, a central bank digital currency would provide no improvement over physical notes and could be worse than current noncash funds transfer systems, especially for a digital currency that could circulate worldwide. In addition, putting a central bank currency in digital form could make it a very attractive target for cyberattacks by giving threat actors a prominent platform on which to focus their efforts. Any implementation would need to adequately deal with a variety of cyber threats--especially for a reserve currency like the U.S. dollar. Second, the issuance of central bank digital currency could have implications for retail banking beyond payments. If a successful central bank digital currency were to become widely used, it could become a substitute for retail banking deposits. This could restrict banks’ ability to make loans for productive economic activities and have broader macroeconomic consequences. Moreover, the parallel coexistence of central bank digital currency with retail banking deposits 6 See, for example, the recent joint Committee on Payments and Market Infrastructures and Markets Committee report “Central Bank Digital Currencies,” March 2018, https://www.bis.org/cpmi/publ/d174.pdf. A Fed-issued digital currency might have implications for the rates and terms of funding for U.S. financial institutions and even the U.S. government as well as the transmission of monetary policy that I will not discuss here. -6could raise the risk of runs on the banking system in times of stress and so have adverse implications for financial stability. Finally, there is no compelling demonstrated need for a Fed-issued digital currency. Most consumers and businesses in the U.S. already make retail payments electronically using debit and credit cards, payment applications, and the automated clearinghouse network. Moreover, people are finding easy ways to make digital payments directly to other people through a variety of mobile apps. New private-sector real-time payments solutions are beginning to gain acceptance in the United States. And the Faster Payments Task Force has laid out a roadmap embraced by a variety of stakeholders for a fast, ubiquitous, and secure payments system to be in place in the United States in the next few years. 7 In short, a multiplicity of mechanisms are likely to be available for American consumers to make payments electronically in real time. As such, it is not obvious what additional value a Fed-issued digital currency would provide over and above these options. Wholesale Digital Settlement Tokens It is important for the Fed and other central banks to continue to research these issues as technology evolves, exploring the technical and economic possibilities and limitations of centralbank-issued digital currencies. Even though the case for a digital currency for general use may not be compelling, opportunities for more targeted and restricted use may nonetheless prove to have value. The private sector has been exploring a variety of ways of deploying the underlying technologies of digital assets that are native to a particular wholesale platform, to help to 7 See https://fedpaymentsimprovement.org/faster-payments/path-to-faster-payments/. -7facilitate finality of settlement. Such wholesale digital settlement tokens could potentially reduce the time and costs required for wholesale financial transactions. This is being discussed, for instance, for the use cases of interbank payments, securities settlements, and cross-border transactions, where the introduction of a digital token native to a platform may facilitate certain types of settlement. Likewise, it is possible at some point in the future that a limited central bank digital instrument that serves as a settlement asset for wholesale payment and settlement activity may hold some promise. Several central banks have been studying this issue, and we have been actively watching these developments. 8 We are also interested in work that decouples the underlying distributed ledger technology from cryptocurrencies and attempts to build on the benefits of the technology, a topic to which I now turn. Distributed Ledger Technology Even if cryptocurrencies prove to have a very limited role in the future, the technology behind them is likely to live on and offer improvements in the way we transfer and record more traditional financial assets. Distributed ledger technology could also facilitate other applications that could improve the way we share information, validate possessions, and handle logistics. Recall that distributed ledger technology is the mechanism for recordkeeping and transfer of ownership that underpins cryptocurrencies. Over the past few years, the financial industry has conducted a great deal of research and development on how to adapt the more promising aspects of distributed ledger technology for use with more traditional financial assets. The industry has moved a number of these projects through a series of phases, often developing more incremental changes at first in order to gain confidence in the technology before tackling large projects with 8 For example, see Bank of Canada’s Project Jasper, https://www.bankofcanada.ca/research/digital-currencies-andfintech/fintech-experiments-and-projects/. -8significant operational impacts. The industry is making steady progress and some projects could be live in some form this year. Many of the use cases focus on the areas of post-trade clearing and settlement of securities transactions, cross-border payments solutions, and trade finance. The common thread running through these use cases is the presence of operational “pain points” that generate inefficiencies and delays for users. For example, post-trade reconciliation of securities transactions can be a time-consuming and resource-intensive process that involves numerous parties, operational steps, and message flows across the counterparties and their various agents involved in the transactions. Distributed ledger technology has the potential to provide synchronized, real-time views for those counterparties and agents that can speed up the process and reduce errors. For cross-border transactions, the process for sending payments via the existing correspondent banking network can add time and money. Distributed ledger technology could potentially lower the costs and time it takes funds to reach the recipient through more direct connections, reducing the number of intermediaries required to effect the transaction. The financial industry has been working on versions of distributed ledger technology that help address a number of concerns, including the loose governance around the maintenance, security, and reliability of the technology for cryptocurrencies. Most projects are organized either as partnerships between technology and financial services firms or through consortia of technology firms, financial firms, and other interested parties. To some degree, these alliances may provide prototype governance arrangements for future technology deployments in financial services. In addition, there are exchanges and clearinghouses that are actively exploring the use of distributed ledger technology, which represent the more traditional model of multilateral -9organization in the financial markets. Although the governance arrangements may need to evolve over time, one thing that is clear is that strong governance arrangements will be required to provide the coordinated operational and financial risk management for the critical clearing and settlement operations that underpin our financial markets. In addition, the industry continues to make progress on the ability of distributed ledger technology to handle the very large volumes of transactions that take place both in financial markets and in retail payments every day. As I highlighted in 2016, this technical challenge of achieving the necessary scale and through put is an important hurdle. 9 Much of this challenge has been tied to the time it takes to achieve “consensus” on a distributed ledger. Consensus is the process by which new transactions are broadcast to all the participants, or nodes, in the network and each node accepts those new transactions as valid additions to the ledger. The initial consensus method used by Bitcoin, called “proof of work,” is designed to deal with the lack of information and trust among the users of the network by providing tools and incentives to overcome this problem. But it is a highly resource-intensive process that limits the number of transactions that can be processed each second. The proof of work consensus model represents a tradeoff between operational efficiency and scalability, on the one hand, and the ability to operate without sufficient trust or information about the entities in the network, on the other hand. Fundamentally, however, the financial industry does not operate as a trustless network. Rather, the industry has long specialized in the collection and analysis of information about 9 Lael Brainard, “The Use of Distributed Ledger Technologies in Payment, Clearing, and Settlement” (speech delivered at the Institute of International Finance Blockchain Roundtable, Washington, DC, April 14, 2016), https://www.federalreserve.gov/newsevents/speech/brainard20160414a.htm; and Lael Brainard, “Distributed Ledger Technology: Implications for Payments, Clearing, and Settlement” (speech delivered at the Institute of International Finance Annual Meeting Panel on Blockchain, Washington, DC, October 7, 2016), https://www.federalreserve.gov/newsevents/speech/brainard20161007a.htm. - 10 customers and counterparties as a core part of banking operations. Even allowing for the inevitable imperfect information that may result, it would seem natural for the financial industry to be able to leverage institutional information and trust in ways that allow for more efficient methods to achieve consensus than proof of work. Consequently, the industry and the academic community have focused a great deal of attention on various consensus methods that can provide greater scalability either by leveraging trust, which relaxes some operational and incentive constraints, or possibly by devising methods without trust that are much less resource intensive. Some of the technology firms working with the financial industry are taking different approaches in this fast-moving arena. Another important challenge for the industry has been leveraging distributed ledger technology while preserving the confidentiality of transactional information. At its core, distributed ledger technology is a shared ledger across multiple nodes in a network, likely representing multiple firms and legal entities. Ownership records and transactions flows from accounts on such a ledger are typically copied and stored on all the nodes in the network. The financial industry, however, must develop distributed ledgers that adhere to laws, regulations, and policies that protect important information of the parties and their customers. Clearly, a model where every entity on the network can see everyone else’s account holdings and transactions history will not satisfy broad industry confidentiality requirements. In addition, stored data that may be protected cryptographically today may not be protected as the technology continues to advance, which adds even more difficulty and urgency to the work on confidentiality. The industry has been working to develop approaches to preserve confidentiality so that only the authorized parties relevant to a transaction can see the details recorded on the ledger. - 11 Some of these approaches involve encrypting data on the ledger so that the ledgers can still be copied across all the nodes in the network, but an entity cannot look at any element of that ledger except for transactions in which it has been involved. Other approaches include so-called zeroknowledge proofs or ring signatures that allow entities to validate transactions without seeing confidential information. Still others are looking at platforms that connect multiple ledgers rather than having one single ledger that is copied across all nodes in the network. While questions remain about the usefulness and viability of each of these approaches, it is important to underscore that preserving confidentiality is an important area of research. Finally, perhaps the biggest potential benefit for payments, clearing, and settlement of distributed ledger technology may be resiliency. Distributed ledger technology may enable a network to continue to operate even if some of the nodes on the network are compromised because of the ability of the other nodes in the network to pick up the slack and continue processing transactions. One challenge going forward will be to understand the implications that the confidentiality tools and different approaches to consensus under consideration may have on the resilience of the distributed ledger. Given that resiliency is a key potential benefit of distributed ledger technology over existing platforms, it is critical to understand the trade-offs between resiliency and a consensus method that focuses on operational speed, or between resilience and confidentiality. Conclusion It is an exciting time for the financial sector as digital innovations are challenging conventional thinking about currency, money, and payments. Cryptocurrencies are strikingly innovative but also pose challenges associated with speculative dynamics, investor and consumer protections, and money-laundering risks. Although central bank digital currencies may be able - 12 to overcome some of the particular vulnerabilities that cryptocurrencies face, they too have significant challenges related to cybersecurity, money laundering, and the retail financial system. Even so, digital tokens for wholesale payments and some aspects of distributed ledger technology--the key technologies underlying cryptocurrencies--may hold promise for strengthening traditional financial instruments and markets. I have highlighted a few key areas where the technology is advancing to deal with some important policy, business, and operational challenges. The Federal Reserve is dedicated to continuing to monitor industry developments and conduct research in these vital areas. I remain optimistic that the financial sector will find valuable ways to employ distributed ledger technology in the area of payments, clearing, and settlement in coming years. THE FUTURE OF MONEY: DIGITAL CURRENCY HEARING BEFORE THE SUBCOMMITTEE ON MONETARY POLICY AND TRADE OF THE COMMITTEE ON FINANCIAL SERVICES U.S. HOUSE OF REPRESENTATIVES ONE HUNDRED FIFTEENTH CONGRESS SECOND SESSION JULY 18, 2018 Printed for the use of the Committee on Financial Services Serial No. 115–111 ( U.S. GOVERNMENT PUBLISHING OFFICE WASHINGTON nshattuck on FSR297 with DISTILLER 31–510 PDF VerDate Sep 11 2014 07:37 Nov 15, 2018 Jkt 000000 PO 00000 Frm 00001 Fmt 5011 : 2018 Sfmt 5011 G:\GPO PRINTING\DOCS\115TH HEARINGS - 2ND SESSION 2018\2018-07-18 MPT DIGIT HOUSE COMMITTEE ON FINANCIAL SERVICES JEB HENSARLING, Texas, Chairman PATRICK T. MCHENRY, North Carolina, Vice Chairman PETER T. KING, New York EDWARD R. ROYCE, California FRANK D. LUCAS, Oklahoma STEVAN PEARCE, New Mexico BILL POSEY, Florida BLAINE LUETKEMEYER, Missouri BILL HUIZENGA, Michigan SEAN P. DUFFY, Wisconsin STEVE STIVERS, Ohio RANDY HULTGREN, Illinois DENNIS A. ROSS, Florida ROBERT PITTENGER, North Carolina ANN WAGNER, Missouri ANDY BARR, Kentucky KEITH J. ROTHFUS, Pennsylvania LUKE MESSER, Indiana SCOTT TIPTON, Colorado ROGER WILLIAMS, Texas BRUCE POLIQUIN, Maine MIA LOVE, Utah FRENCH HILL, Arkansas TOM EMMER, Minnesota LEE M. ZELDIN, New York DAVID A. TROTT, Michigan BARRY LOUDERMILK, Georgia ALEXANDER X. MOONEY, West Virginia THOMAS MACARTHUR, New Jersey WARREN DAVIDSON, Ohio TED BUDD, North Carolina DAVID KUSTOFF, Tennessee CLAUDIA TENNEY, New York TREY HOLLINGSWORTH, Indiana MAXINE WATERS, California, Ranking Member CAROLYN B. MALONEY, New York NYDIA M. VELÁZQUEZ, New York BRAD SHERMAN, California GREGORY W. MEEKS, New York MICHAEL E. CAPUANO, Massachusetts WM. LACY CLAY, Missouri STEPHEN F. LYNCH, Massachusetts DAVID SCOTT, Georgia AL GREEN, Texas EMANUEL CLEAVER, Missouri GWEN MOORE, Wisconsin KEITH ELLISON, Minnesota ED PERLMUTTER, Colorado JAMES A. HIMES, Connecticut BILL FOSTER, Illinois DANIEL T. KILDEE, Michigan JOHN K. DELANEY, Maryland KYRSTEN SINEMA, Arizona JOYCE BEATTY, Ohio DENNY HECK, Washington JUAN VARGAS, California JOSH GOTTHEIMER, New Jersey VICENTE GONZALEZ, Texas CHARLIE CRIST, Florida RUBEN KIHUEN, Nevada SHANNON MCGAHN, Staff Director nshattuck on FSR297 with DISTILLER (II) VerDate Sep 11 2014 07:37 Nov 15, 2018 Jkt 000000 PO 00000 Frm 00002 Fmt 5904 Sfmt 5904 G:\GPO PRINTING\DOCS\115TH HEARINGS - 2ND SESSION 2018\2018-07-18 MPT DIGIT SUBCOMMITTEE ON MONETARY POLICY AND TRADE ANDY BARR, Kentucky, Chairman ROGER WILLIAMS, Texas, Vice Chairman FRANK D. LUCAS, Oklahoma BILL HUIZENGA, Michigan ROBERT PITTENGER, North Carolina MIA LOVE, Utah FRENCH HILL, Arkansas TOM EMMER, Minnesota ALEXANDER X. MOONEY, West Virginia WARREN DAVIDSON, Ohio CLAUDIA TENNEY, New York TREY HOLLINGSWORTH, Indiana GWEN MOORE, Wisconsin, Ranking Member GREGORY W. MEEKS, New York BILL FOSTER, Illinois BRAD SHERMAN, California AL GREEN, Texas DENNY HECK, Washington DANIEL T. KILDEE, Michigan JUAN VARGAS, California CHARLIE CRIST, Florida nshattuck on FSR297 with DISTILLER (III) VerDate Sep 11 2014 07:37 Nov 15, 2018 Jkt 000000 PO 00000 Frm 00003 Fmt 5904 Sfmt 5904 G:\GPO PRINTING\DOCS\115TH HEARINGS - 2ND SESSION 2018\2018-07-18 MPT DIGIT nshattuck on FSR297 with DISTILLER VerDate Sep 11 2014 07:37 Nov 15, 2018 Jkt 000000 PO 00000 Frm 00004 Fmt 5904 Sfmt 5904 G:\GPO PRINTING\DOCS\115TH HEARINGS - 2ND SESSION 2018\2018-07-18 MPT DIGIT CONTENTS Page Hearing held on: July 18, 2018 ..................................................................................................... Appendix: July 18, 2018 ..................................................................................................... 1 23 WITNESSES WEDNESDAY, JULY 18, 2018 Garratt, Rodney J., Maxwell C. and Mary Pellish Chair, Professor of Economics, University of California Santa Barbara ...................................................... Michel, Norbert J., Director, Center for Data Analysis, The Heritage Foundation ........................................................................................................................ Pollock, Alex J., Distinguished Senior Fellow, R Street Institute ....................... Prasad, Eswar S., Tolani Senior Professor of Trade Policy, Cornell University ................................................................................................ 5 7 10 9 APPENDIX Prepared statements: Garratt, Rodney J. ............................................................................................ Michel, Norbert J. ............................................................................................. Pollock, Alex J. ................................................................................................. Prasad, Eswar S. .............................................................................................. 24 30 39 44 nshattuck on FSR297 with DISTILLER (V) VerDate Sep 11 2014 07:37 Nov 15, 2018 Jkt 000000 PO 00000 Frm 00005 Fmt 5904 Sfmt 5904 G:\GPO PRINTING\DOCS\115TH HEARINGS - 2ND SESSION 2018\2018-07-18 MPT DIGIT APPENDIX July 18, 2018 nshattuck on FSR297 with DISTILLER (23) VerDate Sep 11 2014 07:37 Nov 15, 2018 Jkt 000000 PO 00000 Frm 00029 Fmt 6601 Sfmt 6601 G:\GPO PRINTING\DOCS\115TH HEARINGS - 2ND SESSION 2018\2018-07-18 MPT DIGIT VerDate Sep 11 2014 07:37 Nov 15, 2018 Jkt 000000 PO 00000 Frm 00050 Fmt 6601 Sfmt 6601 Insert offset folio 21 here 31510.021 nshattuck on FSR297 with DISTILLER 44 G:\GPO PRINTING\DOCS\115TH HEARINGS - 2ND SESSION 2018\2018-07-18 MPT DIGIT VerDate Sep 11 2014 07:37 Nov 15, 2018 Jkt 000000 PO 00000 Frm 00067 Fmt 6601 Sfmt 6601 Insert offset folio 38 here 31510.038 nshattuck on FSR297 with DISTILLER 61 G:\GPO PRINTING\DOCS\115TH HEARINGS - 2ND SESSION 2018\2018-07-18 MPT DIGIT 62 VerDate Sep 11 2014 07:37 Nov 15, 2018 Jkt 000000 PO 00000 Frm 00068 Fmt 6601 Sfmt 6011 Insert offset folio 39 here 31510.039 nshattuck on FSR297 with DISTILLER Æ G:\GPO PRINTING\DOCS\115TH HEARINGS - 2ND SESSION 2018\2018-07-18 MPT DIGIT CHAPTER 1.2 THE HISTORY OF U.S. FINANCIAL REGULATION: A THEMATIC OVERVIEW CONTENTS I. INTRODUCTION ................................................................................................ 34 II. THE FIRST AND SECOND BANKS OF THE UNITED STATES (1790–1863) .......... 34 A. The First Bank of the United States ...................................................... 34 B. The Second Bank of the United States .................................................. 36 C. Jackson’s War Against the Second Bank ............................................... 37 D. The Free Banking Era ............................................................................ 38 III. THE RISE OF NATIONAL BANKS (1863–1914) .................................................. 39 A. Uniform National Currency.................................................................... 40 B. The Dual Banking System ...................................................................... 40 C. Unit Banks and Correspondent Banking............................................... 41 D. Interbank Deposits in the Reserve Structure ........................................ 42 IV. THE CREATION OF THE FEDERAL RESERVE SYSTEM (1907–1933) .................. 44 A. The Panic of 1907 .................................................................................... 44 B. Functions of the Federal Reserve System .............................................. 47 C. The Early Federal Reserve System ........................................................ 48 V. THE NEW DEAL AND ITS LEGACY (1933–1960s) ............................................. 49 A. Enhancing Transparency in Securities Markets and Separating Banking from Securities ......................................................................... 50 1. The Securities Act of 1933 and the Exchange Act of 1934 ............. 50 2. The Banking Act of 1933 .................................................................. 51 B. Stabilizing the Banking System Through Deposit Insurance............... 51 VI. A TREND TOWARDS DEREGULATION (1970s–2000s) ....................................... 53 A. Deregulation of Depository Institutions ................................................ 54 B. Loosening of Geographic Restrictions .................................................... 55 C. Re-Regulation of Banks and Thrifts ....................................................... 56 D. The Erosion of the Glass-Steagall Act.................................................... 57 E. Securitization and the New Models of Intermediation ......................... 58 F. The Commodity Futures Modernization Act ......................................... 59 VII. THE FINANCIAL CRISIS AND ITS AFTERMATH .................................................. 59 A. The Financial Crisis ................................................................................ 59 B. The Dodd-Frank Act ............................................................................... 64 C. Contending Visions of the Financial Crisis ........................................... 65 D. Changed Political Climate ...................................................................... 72 33 44 INTRODUCTION TO FINANCIAL REGULATION IV. PART I THE CREATION OF THE FEDERAL RESERVE SYSTEM (1907–1933) Since the demise of the Second Bank, the United States had changed from an agricultural economy to a developing industrial economy. Unlike all other developed nations at the time, however, the United States had not established a lender of last resort. The Federal Reserve System evolved over time. The Panic of 1907 created a strong impetus for reform, and after long and contentious debates, Congress enacted the Federal Reserve Act in 1913. As you will see, the hybrid structure of the Federal Reserve System involved compromises between competing factions, echoing the earlier debates around the First and Second Banks. These compromises were revisited in the wake of the Federal Reserve’s failures in the Great Depression, resulting in significant reforms enacted in 1935. As you read this material, consider the extent to which debates at the time of the Federal Reserve System’s founding remain relevant more than a century later. A. THE PANIC OF 1907 The Panic of 1907, even more so than the string of panics that had preceded it, demonstrated that the U.S. financial system was inherently fragile. The panic began among trust companies that were not regulated as banks but made loans and offered demand deposits. Jon R. Moen & Ellis W. Tallman, The Panic of 1907 FED. RESERVE HISTORY (Dec. 4, 2015) The Panic of 1907 was the first worldwide financial crisis of the twentieth century. It transformed a recession into a contraction surpassed in severity only by the Great Depression. . . . The central role of New York City trust companies distinguishes the Panic of 1907 from earlier panics. Trust companies were state-chartered intermediaries that competed with banks for deposits. . . . [T]hey held a low percentage of cash reserves relative to deposits, around 5 percent, compared with 25 percent for national banks. Because trust-company deposit accounts were demandable in cash, trusts were just as susceptible to runs on deposits as were banks. . . . [T]rusts were large and important to the financial system. Trust companies loaned large sums directly in New York equity markets, including New York Stock Exchange [(NYSE)] brokers. Trusts did not require collateral for these loans, which had to be repaid by the end of the business day. Brokers used these loans to purchase securities for themselves or their clients and then used these securities as collateral for a call loan—an overnight loan that facilitated stock purchases—from a nationally chartered bank. The proceeds of the call loan were used to pay back the initial loan from the trust company. Trusts were a necessary part of this process, because the law prohibited nationally chartered commercial banks from making uncollateralized loans or guaranteeing the payment of checks written by brokers on accounts without sufficient funds. The extra liquidity provided by trusts supported new daily transactions on the floor of the exchange. Runs on trust company deposits, however, short-circuited their role as the initial liquidity provider to the stock market. . . . The parallels between the crises in 1907 and 2008 are striking. During 2007–09, the financial crisis was centered on investment banks, institutions without direct access to CH. 1.2 THE HISTORY OF U.S. FINANCIAL REGULATION: A THEMATIC OVERVIEW the Federal Reserve System. In 1907, widespread depositor withdrawals occurred at New York City trust companies—intermediaries outside the New York Clearing House, the effective lender of last resort. In effect, both financial crises started outside the large banks serving as payments centers. Yet the crises created havoc within markets and among banks that were central to the payments system. Both crises challenged the existing mechanisms used to alleviate crises. The trust companies in 1907 were like the shadow banks in the financial crisis of 2007– 09. Short-term lending during the recent crisis came largely from some shadow banks (hedge funds and money market mutual funds) to fund other shadow banks (investment banks). As key liquidity providers for repurchase agreements, these shadow banks were the “depositors” providing funds for overnight lending to allow investment banks to finance the asset-backed security market, just as uncollateralized loans (overdrafts) by trust companies allowed brokers to purchase stock. Both the trusts and the shadow banks faced runs by their depositors and had to withdraw lending in shortterm credit markets. The Panic of 1907 was initially prompted by market rumor regarding the financial health of an established banking institution. The rumor caused banks around the country to hoard cash and call in their own deposits with other banks, as they feared a bank run. Widespread public distrust ensued, leading to a general demand for deposits, which the banks could not meet, and currency began to disappear from circulation. LUDWIG BENDIX, THE ALDRICH PLAN IN THE LIGHT OF MODERN BANKING 23 (1912). In the midst of the panic, J. Pierpont Morgan, namesake of the firm J.P. Morgan, stepped in and personally pledged his own money to prop up the banking system. Simultaneously, Morgan brought in other bankers and institutions to do the same. In a famous finale to this rescue, Morgan gathered the other financiers in his lavish library and held them there until they agreed to provide the funds necessary to stave off collapse. The Treasury also contributed tens of millions of dollars, but Morgan received the lion’s share of credit for saving the day. Morgan’s involvement quite understandably raised concerns about leaving the fate of the burgeoning financial system in the hands of one private individual. See RON CHERNOW, THE HOUSE OF MORGAN: AN AMERICAN BANKING DYNASTY AND THE RISE OF MODERN FINANCE 126–28 (1990); see also PETER CONTI-BROWN, THE POWER AND INDEPENDENCE OF THE FEDERAL RESERVE 17 (2016); GARY B. GORTON, MISUNDERSTANDING FINANCIAL CRISES: WHY WE DON’T SEE THEM COMING 142–43 (2012). After the Panic of 1907, many supported reforms that could reduce the banking system’s susceptibility to panics as a result of seasonal demands. Consensus for reform did not, however, coalesce in favor of creating a central bank until years afterwards. The period between 1907 and 1913 was marked by vigorous debate. Proponents of decentralized local government, reprising the local power ideals of Jeffersonian and Jacksonian democracy, set themselves against proponents of a strong national government modeled more closely on the Hamiltonian tradition. Before the creation of a single national currency, members of the camp in favor of decentralized local government, led by William Jennings Bryan, favored reforms that would expand the U.S. monetary supply by allowing banks to issue private currency backed not by specie or government 45 46 INTRODUCTION TO FINANCIAL REGULATION PART I securities but by the assets they held, such as loans and discounted notes. Members of the opposing camp, including a German émigré banker named Paul Warburg and, later, Senator Nelson Aldrich, favored the creation of a central bank. Warburg arrived in New York in 1902 and, reared in the efficiently synchronized world of German banking, acutely perceived with an outsider’s eye the flaws in the banking system that myopic Americans had missed: Warburg stressed that a central bank was a requisite for developing deeper, more liquid credit markets. Even in a second tongue, Warburg waxed poetic over the centralized credit systems of Europe, where “the credit of the whole nation—that is, the farmer, merchant and manufacturer . . . becomes available as a means of exchange.” Warburg wanted Americans to see that their system was weakened by its lack of unity. He vividly compared its banks to the infantry in a disorganized platoon. “Instead of sending an army,” he admonished, “we send each soldier to fight alone.” LOWENSTEIN, AMERICA’S BANK: THE EPIC STRUGGLE TO CREATE THE FEDERAL RESERVE, at 55–56. In the wake of the Panic of 1907, Congress established a commission to study the banking systems of Europe and the United States. Eventually, Aldrich and a small number of others who went on to lead the central banking movement held a secret summit on Jekyll Island in 1910, developing a plan, which came to be called the Aldrich Plan, setting out their vision for a U.S. central bank. The Aldrich Plan remained a closely guarded secret, however, even as the Jekyll Island group began to unfold their designs in the face of intense opposition to the idea of a central bank. See id. at 107–23. The election of progressive Democrat Woodrow Wilson in 1912 sparked new impetus for reform. Fearing political repercussions from rural voters reflexively opposed to concentrated financial power in any form, however, Wilson avoided publicly supporting the establishment of a central bank during his 1912 presidential campaign but privately signaled he would be receptive. Id.; see also CONTI-BROWN, THE POWER AND INDEPENDENCE OF THE FEDERAL RESERVE; ALLAN H. MELTZER, A HISTORY OF THE FEDERAL RESERVE (2002); JOHN H. WOOD, CENTRAL BANKING IN A DEMOCRACY: THE FEDERAL RESERVE AND ITS ALTERNATIVES (2015); RICHARD H. TIMBERLAKE, THE ORIGINS OF CENTRAL BANKING IN THE UNITED STATES 6 (1978). As President, Wilson favored a strong, publicly run central bank based in Washington. Large banks favored a privately run system based on the Bank of England and British clearinghouse model, while those in the Jeffersonian tradition wanted to disperse power away from Washington. Ultimately, Congress enacted a compromise. Then-representative Carter Glass, a Democrat from Virginia, favored a decentralized system of private reserve banks. In a compromise that became the Federal Reserve Act, Wilson accepted a hybrid public-private structure with a publicly appointed Board of Governors based in Washington that supervised 12 privately owned Federal Reserve Banks, each corresponding to a geographically apportioned district. A version of this compromise still exists today. The uneasy balance between decentralized authority focused on local interests and intermittent crises CH. 1.2 THE HISTORY OF U.S. FINANCIAL REGULATION: A THEMATIC OVERVIEW requiring central action in Washington continues to define the Federal Reserve System, though it resulted in especially deleterious friction in the Federal Reserve Board’s early decades. See CONTI-BROWN, THE POWER AND INDEPENDENCE OF THE FEDERAL RESERVE. B. FUNCTIONS OF THE FEDERAL RESERVE SYSTEM The Federal Reserve Act addressed two major deficiencies in the banking system that the Panic of 1907 exposed: the lack of an elastic currency supply and the absence of a lender of last resort. During panics, banks faced with a shortage of cash had no way to expand the money supply. Issuing more national banknotes to serve as an emergency currency was not an option because banknotes had to be backed by a corresponding value of Treasury bonds. Additional national banknotes could therefore only be issued through an increase in Treasury bonds deposited with the Comptroller of Currency. In times of panic, that was not an option because banks did not have enough money to purchase more Treasury bonds, let alone pay their depositors. In essence, national banknotes were not an elastic currency because their volume could not expand with rising need during a crisis. The Federal Reserve Act solved this problem by creating Federal Reserve notes, legal tender whose supply the Federal Reserve System could increase or decrease as needed. The Federal Reserve Banks were authorized to discount short-term commercial and agricultural paper for their member banks against the proceeds of which Federal Reserve notes could be issued and put in circulation. Then, as the discounted paper was paid off, the Federal Reserve notes would be withdrawn from circulation. In this way, the Federal Reserve could control the amount of Federal Reserve notes in circulation and, thus, the elasticity of the money supply. As the Panic of 1907 demonstrated, the U.S. banking system lacked an adequate lender of last resort that could provide discretionary liquidity to banks in times of crisis. As you will learn in Chapter 9.1, the Federal Reserve Act allowed the Federal Reserve System to serve as a lender of last resort by replenishing member banks’ reserves when banks experienced liquidity shortfalls and could not obtain funds elsewhere. When the banking system was unstable, Federal Reserve Banks could now stand ready to inject funds into fundamentally solvent banks experiencing temporary liquidity problems. By doing so, the Federal Reserve System could help these banks avoid forced asset sales to meet their obligations and also discourage depositors from running on a bank due to concerns over a bank’s liquidity. Aldrich and his allies hoped that this lender of last resort function would mitigate the risk of bank failure and contagion. The concept of the lender of last resort dates back to 1797, when it was first suggested by Sir Francis Baring that the Bank of England should be the “dernier resort.” SIR FRANCIS BARING, OBSERVATIONS ON THE ESTABLISHMENT OF THE BANK OF ENGLAND AND ON THE PAPER CIRCULATION OF THE COUNTRY 22 (Augustus M. Kelley ed., 1967). In reality, however, the Bank of England did not consistently provide liquidity in times of crisis, nor did it embrace its role as lender of last resort until the mid-19th Century. Thomas M. Humphrey & Robert 47 48 INTRODUCTION TO FINANCIAL REGULATION PART I E. Keleher, The Lender of Last Resort: A Historical Perspective, 4 CATO J. 275, 299–300 (1984). The Bank of England was persuaded to adopt the role of lender of last resort by Walter Bagehot. Id. at 291–305. Bagehot argued that, in times of panic, the Bank of England should lend freely in accordance with two rules. First, the loans should only be made at high rates of interest, known as penalty rates. Second, the Bank of England should make loans against all collateral that was considered good, commonly pledged, and easily convertible during ordinary times, so as not to exacerbate the panic. WALTER BAGEHOT, LOMBARD STREET: A DESCRIPTION OF THE MONEY MARKET, ch. 7 ¶¶ 57–60 (Henry S. King & Co. 1873). C. THE EARLY FEDERAL RESERVE SYSTEM The Federal Reserve System started operations in November 1914, just after the beginning of World War I. Contending with a more or less constant power struggle between the regional Federal Reserve Banks and the Board in Washington, the Federal Reserve System, by tightening credit during the Stock Market Crash of 1929, worsened the Great Depression and contributed to the failure of many banks between 1929 and 1933. See LIAQUAT AHAMED, LORDS OF FINANCE: THE BANKERS WHO BROKE THE WORLD 501–02 (2009); see also BEN S. BERNANKE, ESSAYS ON THE GREAT DEPRESSION (2000); MILTON FRIEDMAN & ANNA J. SCHWARTZ, THE GREAT CONTRACTION, 1929–1933 (1969). Twenty years after the passage of the Federal Reserve Act, the Banking Act of 1935 altered the Federal Reserve System’s structure and governance. The Act was the brainchild of Marriner S. Eccles, special assistant to the Secretary of the Treasury. See CONTI-BROWN, THE POWER AND INDEPENDENCE OF THE FEDERAL RESERVE. Eccles’s goal was to make the Washington-based Federal Reserve Board the true center of national banking policy and to abolish the Reserve Banks, which were dominated by regional banking interests. Whereas in 1913, limiting federal power was a principal concern, in 1935, Eccles’s principal focus was increasing the power of the Washington-based Federal Reserve Board. Id; see also LOWENSTEIN, AMERICA’S BANK: THE EPIC STRUGGLE TO CREATE THE FEDERAL RESERVE. Eccles’s vision was partially embodied in the Banking Act of 1935. The act strengthened the Federal Reserve Board and prevented the Federal Reserve Banks from undermining a national interest rate and monetary policy; the Federal Reserve Banks, however, were permitted to participate in the Federal Open Market Committee (FOMC), through which the Federal Reserve System sets monetary policy. According to one account, the Federal Reserve System was transformed from “private banks running a private banking policy with public benefits to a public central bank in the modern sense of the word.” CONTI-BROWN, THE POWER AND INDEPENDENCE OF THE FEDERAL RESERVE, at 31. ——————— 1. The Federal Reserve Board’s Evolving Role. The Federal Reserve Board’s role in financial regulation has continued to expand. For instance, in the Bank Holding Company Act of 1956 (BHCA), Congress gave the Federal Reserve Board supervisory and regulatory power over bank holding companies (BHCs). We will discuss BHCs in Chapter 6.1. Twenty years later, amidst high inflation CH. 1.2 THE HISTORY OF U.S. FINANCIAL REGULATION: A THEMATIC OVERVIEW and unemployment, Congress charged the Federal Reserve Board with pursuing the goals of “maximum employment, stable prices, and moderate long-term interest rates.” 12 U.S.C. § 225a (1977) (2012). After the Financial Crisis, Congress put the Federal Reserve Board in charge of supervising financial companies that are deemed systemically important. See, e.g., DoddFrank Act § 113. The Federal Reserve Board’s power, however, remains controversial, and many critics want the Federal Reserve Board weakened or even abolished. Critics allege that the Federal Reserve Board is secretive and undemocratic, and they argue that the Federal Reserve Board’s response to the Financial Crisis exceeded its statutory powers. See, e.g., RON PAUL, END THE FED (2009); see also ROTHBARD, THE CASE AGAINST THE FED. You will learn about the Federal Reserve Board in Chapter 9.1. ——————— V. THE NEW DEAL AND ITS LEGACY (1933–1960s) Following the end of World War I and into the 1920s, the United States enjoyed unprecedented prosperity. Some 20 million shareholders took advantage of post-war prosperity and invested in the stock market, fueling its spectacular growth. This growth, combined with a flow of loose credit, fueled risky and speculative activities. Groups of wealthy investors traded large pools of stocks to manipulate prices and enjoy speculative windfalls. Commercial banks used depositors’ funds, which at the time were uninsured, to make speculative loans. Small investors acquired stock beyond their means from brokers who required only a small down payment, or margin, while loaning the remaining price of the stock. Fueling the fire, securities regulation was either vastly inadequate or nonexistent. Although many states attempted to regulate the securities markets at a local level through Blue Sky laws, brokers and dealers found it easy to evade this state regulation. When stock prices began to decline in September 1929, brokers started demanding repayment of their loans. Many investors lacked sufficient cash to satisfy their obligations, and they resorted to selling their stock. These fire sales further depressed prices, creating a vicious cycle that soon collapsed, bringing the country’s financial system to disaster. On October 24, 1929, dubbed Black Thursday, a selling frenzy resulted in record one-day stock market losses that totaled $9 billion. The following Tuesday, October 29, the market traded 16 million shares, and by November stocks shed some $26 billion in value. The losses did not end there: Between September 1, 1929, and July 1, 1932, the value of all stocks listed on the [NYSE] shrank from a total of nearly $90 billion to just under $16 billion—a loss of 83 percent. In a comparable period, bonds listed on the [NYSE] declined from a value of $49 billion to $31 billion. . . . Nor did these figures, staggering as they were, fully gauge the extent of the 1929–1932 stock market crash. During the post-World War I decade, approximately $50 billion of new securities were sold in the United States. Approximately half or $25 billion would prove near or totally valueless. Leading “blue-chip” securities, including General Electric, Sears, Roebuck, and U.S. Steel common 49 CHAPTER 9.1 LENDER OF LAST RESORT CONTENTS I. INTRODUCTION .............................................................................................. 935 II. HISTORY ........................................................................................................ 936 A. The Bank of England ............................................................................ 936 B. The United States Without a Lender of Last Resort ........................... 938 C. Today’s Federal Reserve ....................................................................... 939 III. TRADITIONAL LENDER OF LAST RESORT ....................................................... 939 A. Open Market Operations ...................................................................... 941 B. The Discount Window ........................................................................... 942 C. The Problem of Stigma and the Early Financial Crisis ...................... 944 IV. FEDERAL RESERVE EMERGENCY LENDING: SECTION 13(3) .......................... 947 A. The History of Section 13(3) ................................................................. 948 B. Aftermath and Reform .......................................................................... 951 C. The Lender of Last Resort Debate ....................................................... 953 1. Central Bank Independence ........................................................... 953 2. Criticisms of the Federal Reserve .................................................. 955 3. The Future of Lender of Last Resort .............................................. 957 V. THE FEDERAL RESERVE AS THE INTERNATIONAL LENDER OF LAST RESORT? ............................................................................................... 959 I. INTRODUCTION In this Chapter, we take a closer look at the central bank and its function as the lender of last resort. In the United States, the Federal Reserve System plays this role. Deposit insurance and the lender of last resort are often discussed as complementary since they both protect against runs, but the lender of last resort is a 19th Century idea and older than deposit insurance plans, which only developed in the early 20th Century. Years of contentious debate accompanied the creation of the Federal Reserve System at the beginning of the 20th Century and, later, the controversy over whether the Federal Reserve used its lender of last resort authority aggressively enough during the 1930s. During the intervening decades of relative financial stability in the United States, however, intellectual and political interest in the function of the lender of last resort waned. As Professor Peter Conti-Brown writes, until roughly the early 1960s, “Central banking was in the hinterland; fiscal policy—the stuff of taxes and budgets and spending and deficits—was at 935 936 LENDER OF LAST RESORT AND RESOLUTION PART IX the core.” PETER CONTI-BROWN, THE POWER AND INDEPENDENCE OF THE FEDERAL RESERVE ix (2016). This lack of attention changed abruptly with the Financial Crisis, when many central banks, including the Federal Reserve, made heavy and creative use of their lender of last resort powers to avert a global financial catastrophe. Earlier crises had prompted assertive but incremental action. For example, after Thailand’s sharp devaluation of the baht in July 1997, central banks established currency swap lines to facilitate foreign exchange and alleviate the market disruption that became known as the Asian financial crisis. The Financial Crisis pushed the Federal Reserve and other lenders of last resort to act far more aggressively. In a highly volatile market, a financial conglomerate might remain solvent while in desperate need of liquidity, but it might already have crossed the brink. Many categorical distinctions—insolvency versus illiquidity; banking versus non-banking activities; systemic versus idiosyncratic risks— collapsed under the weight of overwhelming uncertainty. Clear-eyed theory often gave way to more nuanced and fraught experimentation as the Federal Reserve Board and other policy-makers struggled to understand and respond to the Financial Crisis. We begin with a history of the lender of last resort, starting with the Bank of England. We then discuss the traditional lender of last resort power in the United States, as well as the extraordinary use of the lender of last resort powers during the Financial Crisis. The Financial Crisis emergency lending sparked a vigorous academic and political debate over the proper role of the lender of last resort, which we will try to capture. The Chapter concludes with an examination of whether the Federal Reserve has become a truly international lender of last resort. II. A. HISTORY THE BANK OF ENGLAND The Bank of England has served as the model for other central bank lenders of last resort. It was granted a corporate charter in 1694 as a private company after a crushing naval defeat and the threat of French invasion required the English crown to rebuild its navy. The English crown had neither funds nor good credit, so it needed another means to get a loan. In order to entice subscribers for the large loan, the crown agreed to grant prospective lenders a bank charter. ANDREAS M. ANDREADÉS, HISTORY OF THE BANK OF ENGLAND 54–59 (4th ed. 2013). The charter granted the Bank of England’s private shareholders limited liability protection, which at the time was an innovative idea, especially in banking. Id. at 83. The Bank of England was initially the government’s preferred bank, holding its balances and lending it money. CHARLES GOODHART, THE EVOLUTION OF CENTRAL BANKS 5 (3d ed. 1981). It was also empowered to issue banknotes against the government’s debt and deal in bills of exchange and gold or silver bullion. ANDREAS M. ANDREADÉS, HISTORY OF THE BANK OF ENGLAND, at 73. Its charter had to be renewed every 20 years and its powers expanded or contracted according to the politics of the time. The 20-year charter, CH. 9.1 LENDER OF LAST RESORT a common feature historically, and private ownership were to be repeated in the earliest, ultimately doomed, attempts to establish a central bank in the United States. See Chapter 1.2. In 1844, the Bank Charter Act granted the Bank of England a monopoly, with certain exceptions, on issuing banknotes, and these notes were declared legal tender. Thomas M. Humphrey & Robert E. Keleher, The Lender of Last Resort: A Historical Perspective, 4 CATO J. 275, 289–97 (Spring/Summer 1984); see also ANDREADÉS, HISTORY OF THE BANK OF ENGLAND, at 188–89, 200–01, 288–91. It is unclear when the Bank of England first acted as a lender of last resort. The concept of the lender of last resort was first suggested by Sir Francis Baring in 1797, when he wrote that the Bank of England was the “dernier ressort.” SIR FRANCIS BARING, OBSERVATIONS ON THE ESTABLISHMENT OF THE BANK OF ENGLAND AND ON THE PAPER CIRCULATION OF THE COUNTRY 22 (Augustus M. Kelley 1967); see also Humphrey & Keleher, The Lender of Last Resort: A Historical Perspective, at 282. Baring was not describing a self-conscious role the bank was performing, but rather the brute fact that the Bank of England was the last one left after every other lender had turned down a request for a loan. Richard S. Grossman & Hugh Rockoff, Fighting the Last War: Economists on the Lender of Last Resort, in CENTRAL BANKS AT A CROSSROADS: WHAT CAN WE LEARN FROM HISTORY?, 231, 245 (Michael D. Bordo et al. eds., 2016). The Bank of England maintained the official stance that it was “no more bound to support commercial credit than any other bankers are” and did not consistently provide liquidity in times of crisis. David Kynaston, TILL TIME’S LAST SAND: A HISTORY OF THE BANK OF ENGLAND 1694–2013, at 152 (2017). The Bank of England’s role in the financial system was nonetheless widely acknowledged by the mid-19th Century, and an 1848 Parliamentary report observed that “the Bank [was subject to] the duty of a consideration of the public interest, not indeed enacted or defined by law, but which Parliament in its various transactions with the Bank has always recognized and which the Bank has never disclaimed.” Id. In response to the inconsistent policies of the Bank of England when confronting panics during the 19th Century, Walter Bagehot wrote LOMBARD STREET: A DESCRIPTION OF THE MONEY MARKET. Although the Bank of England had sometimes supplied some liquidity to help alleviate panics, it had not developed a consistent policy. Building on the earlier, seminal work of Henry Thornton, Bagehot’s book was a landmark, at once sophisticated and accessible, arguing persuasively that the Bank of England’s actions during an earlier panic in 1825 defined the exemplary lender of last resort. See HENRY THORNTON, AN ENQUIRY INTO THE NATURE AND EFFECTS OF THE PAPER CREDIT OF GREAT BRITAIN (1802); Humphrey & Keleher, The Lender of Last Resort: A Historical Perspective, at 297–305. Walter Bagehot, Lombard Street: A Description of the Money Market 57–60 (1873) And with the Bank of England, as with other Banks in the same case, these advances, if they are to be made at all, should be made so as if possible to obtain the object for which they are made. The end is to stay the panic; and the advances should, if possible, stay the panic. And for this purpose there are two rules: 937 938 LENDER OF LAST RESORT AND RESOLUTION PART IX First, That these loans should only be made at a very high rate of interest. This will operate as a heavy fine on unreasonable timidity, and will prevent the greatest number of applications by persons who do not require it. The rate should be raised early in the panic, so that the fine may be paid early; that no one may borrow out of idle precaution without paying well for it; that the Banking reserve may be protected as far as possible. Secondly. That at this rate these advances should be made on all good banking securities, and as largely as the public ask for them. The reason is plain. The object is to stay alarm, and nothing therefore should be done to cause alarm. But the way to cause alarm is to refuse someone who has good security to offer. . . . If it is known that the Bank of England is freely advancing on what in ordinary times is reckoned a good security—on what is then commonly pledged and easily convertible—the alarm of the solvent merchants and bankers will be stayed. But if securities, really good and usually convertible, are refused by the Bank, the alarm will not abate, the other loans made will fail in obtaining their end, and the panic will become worse and worse. . . . The only safe plan for the Bank is the brave plan, to lend in a panic on every kind of current security, or every sort on which money is ordinarily and usually lent. This policy may not save the Bank; but if it do not, nothing will save it. For an excellent review that places Bagehot’s book in context, see Peter Conti-Brown, Misreading Walter Bagehot: What Lombard Street Really Means for Central Banking¸ NEW RAMBLER (Dec. 14, 2015). Bagehot’s dictum, the name by which the modern interpretation of his argument is commonly known, asserts that, in times of panic, the Bank of England should lend freely against what is considered to be good collateral in normal times at high interest rates. Modern articulations of Bagehot’s dictum also include the condition that central banks should lend only to solvent borrowers, a point that would have been implicit when Bagehot wrote but, as we discuss later, has additional relevance today. The actions of the Bank of England reduced the number of 19th Century banking panics in England, especially as compared to the United States, which, after the first and second Banks of the United States’ charters lapsed, did not have a central bank to serve as a lender of last resort. See Chapter 1.2. B. THE UNITED STATES WITHOUT A LENDER OF LAST RESORT At the turn of the 20th Century, the United States was the only developed economy without a lender of last resort. In the absence of a central bank, private clearinghouses stepped in to fill that role. In 1853, certain New York banks joined together to form the New York Clearing House Association. During banking panics, when New York banks became illiquid, the New York Clearing House issued clearinghouse certificates against the banknotes of its members. Member banks paid interest on the certificates and accepted them as payment, relieving other member banks of the need to settle their transactions in greenbacks or specie. This allowed those other banks to use the greenbacks or specie to satisfy deposit withdrawals, making a collapse less likely. The New York Clearing House served as an ad hoc lender of last resort, injecting liquidity into the New York banks by converting banknotes that had ceased to function as money into clearinghouse certificates that were widely accepted as money, and thus avoiding or mitigating a contraction in the supply of money even during a CH. 9.1 LENDER OF LAST RESORT panic. See NEW YORK CLEARING HOUSE ASSOCIATION RECORDS, 1868–1950; JOHN H. WOOD, CENTRAL BANKING IN A DEMOCRACY: THE FEDERAL RESERVE AND ITS ALTERNATIVES 56–60 (2015). Similar systems developed elsewhere, notably in New England under the Suffolk Bank System. The Suffolk Bank of Boston issued notes in exchange for specie or local banknotes, creating a kind of regional currency between 1824 and 1858. GARY WALTON & HUGH ROCKOFF, HISTORY OF THE AMERICAN ECONOMY 215 (2005); C.J. Maloney, Life Without the Fed: The Suffolk System, MISES INST. (Jan. 5, 2011); see also ROGER LOWENSTEIN, AMERICA’S BANK: THE EPIC STRUGGLE TO CREATE THE FEDERAL RESERVE (2015). The weakness of the clearinghouse system was that its solutions were local and the certificates worked only in the city where they were issued. LOWENSTEIN, AMERICA’S BANK, at 41. As discussed in Chapter 1.2, the severity of the Panic of 1907 was the turning point for the development of a central bank to serve as a lender of last resort in the United States. To address the deficiencies in the banking system that were exposed in that crisis, the Federal Reserve Act was passed in 1913. C. TODAY’S FEDERAL RESERVE Today’s Federal Reserve System retains 12 Federal Reserve Banks that service financial institutions in 12 Federal Reserve districts. These reserve banks act as banker’s banks, providing a wide variety of services such as storing currency and processing checks and electronic payments for both banking institutions and the federal government. The Federal Reserve Banks are overseen by a seven-person Board of Governors based in Washington. Each of these Governors is appointed by the President of the United States and serves a 14-year term. The Chairman of the Board of Governors is appointed from among its members for a four-year term. For a more thorough exploration of the governing structure of the Federal Reserve, see BD. OF THE FED. RESERVE SYS., THE FEDERAL RESERVE SYSTEM: PURPOSES & FUNCTIONS 1–13 (2005). We discussed the Federal Reserve’s supervisory authority with respect to bank holding companies (BHCs) and its role in supervising systemic non-bank financial companies in Chapters 6.1 and 6.3, as well as its role in the payments system and check clearing in Chapter 7.2. The Federal Reserve is the primary organ responsible for carrying out U.S. monetary policy and, for many, that is its most crucial role. III. TRADITIONAL LENDER OF LAST RESORT As lender of last resort, each of the 12 Federal Reserve Banks can replenish banks’ liquidity, on a collateralized basis, when they experience shortfalls and cannot obtain funds elsewhere. See Kathryn Judge, Three Discount Windows, 99 CORNELL L. REV. 795 (2014). When the banking system is stable, private commercial banks can overcome liquidity shortfalls by borrowing from one another in the interbank lending market. When the banking system is unstable, however, banks may hesitate to lend to one another because they are unable to determine which institutions are solvent or sufficiently liquid to withstand a run by their depositors or other short-term creditors. By standing ready to inject funds into fundamentally solvent banks experiencing temporary liquidity 939 940 LENDER OF LAST RESORT AND RESOLUTION PART IX problems, the lender of last resort helps these banks avoid forced asset sales to meet their obligations. It also discourages depositors from running on a bank due to concerns over a bank’s liquidity position, which in turn mitigates the risk of bank failure and contagion. The purpose of the lender of last resort, at least in theory, is only to provide temporary fully secured liquidity to solvent banks and not to provide capital to rescue banks from insolvency. A collateralized liquidity injection is distinguishable in theory from a bailout, in which insolvent banks receive equity or equity-like infusions of capital, the impact of which is to shield the bank’s long- and short-term creditors from losses. Those losses are then redistributed to third parties, like taxpayers, the rationale being that the social costs of the bailout are lower than the social costs of letting the bank fail. See Randall D. Guynn, Are Bailouts Inevitable?, 29 YALE J. ON REG. 121, 125–29 (2012). This section focuses on the Federal Reserve’s traditional lender of last resort function. It begins with an examination of open market operations. While open market operations are often discussed as a distinct function from the lender of last resort, it makes sense to begin the discussion there for two reasons. First, open market operations are the Federal Reserve’s primary tool for regulating liquidity in the banking system. Second, allegations that the Federal Reserve abused its lender of last resort power during the Financial Crisis have sparked efforts to restrain not only the lender of last resort power, but also the Federal Reserve’s open market operations function. The Federal Reserve has vigorously resisted these efforts to curb its most powerful monetary policy tool. This section concludes by looking at the Federal Reserve’s bilateral lending facility, the discount window, which has historically carried out what most would recognize as the lender of last resort power. ——————— 1. The Federal Reserve’s Informational Role. Professor Kathryn Judge draws a distinction between a liquidity shortage caused by an exogenous shock, such as the terrorist attacks of September 11, 2001, and a persistent liquidity shortage of the sort that culminated in the Financial Crisis. Persistent liquidity shortages, Judge argues, are symptoms of deeper market dysfunction that the infusion of liquidity alone may be insufficient to alleviate. She proposes a broader understanding of the role of the lender of last resort in today’s financial system, in which the Federal Reserve exercises its authority as lender of last resort not only to provide liquidity, but also to collect, synthesize, and disseminate the information necessary to neutralize the crippling uncertainty that may otherwise trigger a contagious panic: [C]ontinually pumping new liquidity into a financial system in the midst of a persistent liquidity shortage may increase the fragility of the system and, on its own, is unlikely to resolve the deeper problems causing those liquidity shortages to persist. . . [W]hen facing persistent liquidity shortages, the [Federal Reserve Board] should instead use the leverage it enjoys by virtue of controlling access to liquidity to improve its understanding of the ailments causing the market dysfunction to persist and to help address those underlying issues. It should also use the information to further policies that can CH. 9.1 LENDER OF LAST RESORT help to counter those underlying issues. . . . Providing credible information and working with other policymakers to ensure the overall financial system is sufficiently capitalized are thus among the issues that the [Federal Reserve Board] should prioritize when facing persistent liquidity shortages. Kathryn Judge, The First Year: The Role of a Modern Lender of Last Resort, 116 COLUM. L. REV. 843, 843 (2016). The traditional prudential view differs in important respects; regulators in the past strove to keep prudential information secret in order to ward off bank runs. The 2009 stress tests began to change that view, as the transparency of the stress tests served to reduce market uncertainty. ——————— A. OPEN MARKET OPERATIONS The Federal Reserve’s primary tool for regulating liquidity in the banking system is the use of open market operations. To expand or contract the money supply, the Federal Reserve purchases or sells U.S. Treasuries on the open market. The broker-dealers that transact directly with the Federal Reserve are known as primary dealers. Dietrich Domanski, Richhild Moessner & William Nelson, Central Banks as Lenders of Last Resort: Experiences During the 2007– 2010 Crisis and Lessons for the Future 51 (Bank for Int’l Settlements, Working Paper No. 79, 2014). You will recall that, even during the Glass-Steagall Act era, commercial banks were affiliated with primary dealers that underwrote and dealt in U.S. government securities. If the Federal Reserve wants to increase liquidity (i.e., the money supply), it will cause one of the Federal Reserve Banks to purchase U.S. Treasuries from a primary dealer. To pay for the purchase, the Federal Reserve Bank credits the account of the primary dealer and debits its own account in an equal amount. In effect, the U.S. Treasuries purchased by a Federal Reserve Bank conducting open market operations are paid for using newly created money. As a result, the amount of money circulating in the banking system is increased. This expansion in the money supply, in turn, allows banks to lend to one another and to the economy more freely. Interbank lending is measured by reference to the Federal Funds Rate, which is the rate that banks charge one another for overnight loans on balances held at a Federal Reserve Bank. A lower Federal Funds Rate reflects a decrease in the cost of lending to the economy, typically resulting in an expansion of the supply of both money and credit through a mechanism generally known as the money multiplier. JAMES R. KEARL, ECONOMICS AND PUBLIC POLICY: AN ANALYTICAL APPROACH 422–27, 792 (6th ed. 2011). Conversely, the sale of government securities will have the opposite effect of contracting the money supply, as well as the supply of money and credit throughout the banking system. The Federal Reserve’s course in open market operations is determined by the Federal Open Market Committee (FOMC), which is composed of all seven members of the Board of Governors, the President of the Federal Reserve Bank of New York, and an annually rotating group of Presidents from four other Federal Reserve Banks. 12 U.S.C. § 263 (2012). The Federal Reserve undertook aggressive open market operations to increase liquidity during and after past crises. For example, aggressive open 941 942 LENDER OF LAST RESORT AND RESOLUTION PART IX market operations were key in mitigating broad financial market instability during the 1987 stock market crash. Mark Carlson, A Brief History of the 1987 Stock Market Crash with a Discussion of the Federal Reserve Response (FEDS Working Paper No. 2007–13, 2007). The Federal Funds Rate remained close to 0% for several years until December 2015, meaning that banks could effectively lend to each other interest-free. Bd. of Governors of the Fed. Reserve Sys., Selected Interest Rates (Weekly) (Apr. 6, 2015). Despite the role that open market operations play in improving bank liquidity, it is important to remember that the primary purpose of open market operations is not to regulate the health of banks, but rather to expand or contract the money supply, adjust interest rates, and either stimulate or tamp down the availability of credit. B. THE DISCOUNT WINDOW The fact that a lender of last resort stands ready to exchange cash for illiquid assets during periods of scarce liquidity has many salutary effects. The existence of the lender of last resort discourages depositors from participating in a run. Paul Tucker, The Lender of Last Resort and Modern Central Banking: Principles and Reconstruction (Bank for Int’l Settlements, Working Paper No. 79b, 2014) [hereinafter Tucker, Lender of Last Resort]. In the event that a run does occur, central bank liquidity helps an institution with liquidity difficulties avoid asset fire-sales and insolvency. A central bank is in the best position to act as a lender of last resort for several reasons. Central banks often act as prudential supervisors. This regulatory role positions them to assess banks’ health and avoid lending to an insolvent institution. Judge, Three Discount Windows, at 808. In a world of fiat currency, central banks possess the ability to print or otherwise create central bank money. Id. at 807. Their ability to make emergency loans is not constrained by a lack of money, and they have no incentive to hoard liquid assets in times of financial instability. Id. at 808. Central banks with power to expand the money supply are, uniquely, indifferent to liquidity risk. In normal times, the Federal Reserve provides collateralized loans through its bilateral lending facility, the discount window. The interest rate on discount window loans, known as the discount rate, is set by Federal Reserve Banks in each district, subject to approval by the Board of Governors. Normally, only depository institutions are permitted to access the discount window. 12 C.F.R. § 201.1(b) (2017). Discount window loans contain terms designed to mitigate problems of adverse selection and moral hazard. Banks seeking to tap the discount window must post collateral at the window in exchange for a loan of cash. Judge, Three Discount Windows, at 797. A broad range of high-quality collateral is acceptable, including loans and securities. William Nelson, Lessons from Lender of Last Resort Actions During the Crisis: The Federal Reserve Experience 78 (Bank for Int’l Settlements, Working Paper No. 79d, 2014). If a bank borrower turns out to be insolvent, the Federal Reserve can then sell that collateral for cash to satisfy the bank’s secured obligation. Furthermore, discount window loans come with a haircut, meaning that bank borrowers post collateral that has a higher value than the cash they receive in return. This overcollateralization further protects CH. 9.1 LENDER OF LAST RESORT the Federal Reserve against loss. There is no known instance of a Federal Reserve Bank suffering a loss on a secured loan from the discount window. In addition to these safeguards, all discount window loans carry interest rates higher than those prevailing in the market to discourage banks from using the facility. This above-market rate for federal funds varies depending on the soundness of the banks taking out the loan. Banks in good condition can borrow from the Federal Reserve’s standing lending facility at 100 basis points above the targeted Federal Funds Rate. These loans are called primary credit discount loans, and the interest rate on these loans is called the primary rate. Weaker banks that do not qualify for primary discount loans must take a secondary credit discount loan, which usually carries an interest rate 50 basis points higher than the primary rate. The Federal Reserve exercises greater oversight over institutions that take out secondary credit discount loans. Id. If a bank is “critically undercapitalized” under the prompt corrective action regime, which we explored in Chapter 2.5, the Federal Reserve can only extend a loan to that institution if it matures in five days or less. 12 C.F.R. § 201.5. Finally, the Federal Reserve offers seasonal credit to small banks that have deposit withdrawals that spike in certain seasons, requiring greater reserves. The interest rates on these loans are typically below the primary rate. ——————— 1. The Monetary Policy Toolkit. Open market operations and the discount window are among the most important of the Federal Reserve’s tools, but they are not the only instruments of monetary policy. Federal regulations require banks to maintain minimum reserve balances at Federal Reserve Banks and, like commercial banks, Federal Reserve Banks pay interest on these deposits. The power to adjust the interest rates paid to commercial banks on their required and excess Federal Reserve Bank balances is one of the Federal Reserve’s conventional monetary policy tools. Lowering interest rates on excess balances incentivizes banks to put these funds to use elsewhere, i.e., by making commercial loans instead. Forward guidance is another conventional monetary policy tool; by issuing policy statements that express a commitment to maintaining low interest rates, the Federal Reserve can promote confidence in the availability of credit in the future and thereby encourage economic growth. 2. Unconventional Monetary Policy Tools. During the period of low interest rates and economic inertia that followed the Financial Crisis, the effectiveness of conventional monetary policy tools was constrained by what is known as the zero lower bound problem; further reduction may be impossible when interest rates are already at or near 0%. In theory, reserves might carry a negative interest rate, charging banks for maintaining excess reserves. Some central banks, including the European Central Bank and the Bank of Japan, have ventured into negative interest rate territory. The possibility of the Federal Reserve following the same course remains a subject of academic debate, but the Federal Reserve did begin using other unconventional tools more extensively and creatively after the Financial Crisis to expand the money supply in an effort to stimulate growth. Most prominently, the Federal Reserve launched the first in a series of quantitative easing programs in 2008. Like open market operations, quantitative easing involves the purchase of assets from market participants by 943 944 LENDER OF LAST RESORT AND RESOLUTION PART IX Federal Reserve Banks to increase the amount of liquidity in the financial system. It is considered unconventional in that the assets purchased include not only U.S. Treasuries but also other types of securities, such as mortgage-backed securities issued by government sponsored enterprises. The Federal Reserve has also experimented with the use of overnight reverse repurchase transactions with money market mutual funds (MMFs) as an alternative to purchasing U.S. Treasuries from primary dealers in open market operations to influence the Federal Funds Rate. 3. Shrinking the Federal Reserve’s Balance Sheet. Quantitative easing resulted in a massive expansion of the Federal Reserve’s balance sheet, which grew from less than $900 billion early in 2008 to more than $4.5 trillion at its peak in 2015 and has remained close to that size for years afterwards. See Bd. of Governors of the Fed. Reserve Sys., Credit and Liquidity Programs and the Balance Sheet: Recent Balance Sheet Trends (last updated Mar. 9, 2018) (select the “Total Assets of the Federal Reserve” chart). In September 2017, the FOMC stated that it “will initiate [a] balance sheet normalization program,” using Federal Reserve parlance to announce its plans to reduce the size of the balance sheet over time. Just as quantitative easing supported financial stability and economic growth, there are concerns that the Federal Reserve’s balance sheet reduction may disrupt markets if it proceeds too quickly or too far. Former Federal Reserve Chairman Ben Bernanke supports a prudent course of action, arguing for the benefits of keeping a large balance sheet, such as better transmission of monetary policy, an increased supply of safe assets available to market participants and a greater ability for the Federal Reserve to provide liquidity during a crisis. Furthermore, he argues that the process of reduction should be “passive and predictable,” with the Federal Reserve declining to reinvest the proceeds of assets once they mature but not proactively selling assets off. See Ben S. Bernanke, Shrinking the Fed’s Balance Sheet, BROOKINGS INST.: BEN BERNANKE’S BLOG (Jan. 26, 2017), Calling for more aggressive moves, James Bullard, President of the Federal Reserve Bank of St. Louis, argues that the large balance sheet reflects a monetary policy that is distorting markets and that reducing its size more quickly would “allow for more balance-sheet ‘policy space’ in the future” in case of a recession. James Bullard, A Case for Shrinking the Fed’s Balance Sheet, THE REGIONAL ECONOMIST (Second Quarter 2017). One commentator summarizes the market’s initial reaction to the FOMC’s announcement of balance sheet normalization: “For now, markets will continue to treat the Fed's normalization as the equivalent of watching paint dry: a slow, uneventful and relatively predictable process. Whether such expectations are validated over the longer term, however, is far from clear.” Mohamed A. El-Erian, Opinion, The Next Step in the Federal Reserve’s Beautiful Normalization, BLOOMBERG VIEW (Sept. 20, 2017). ——————— C. THE PROBLEM OF STIGMA AND THE EARLY FINANCIAL CRISIS The discount window was rarely used before the Financial Crisis, primarily due to the problem of stigma. Banks are hesitant to tap the discount window for fear of projecting weakness, and only the most troubled banks would do so. This AS PREPARED FOR DELIVERY Winds of Change: The Case for New Digital Currency By Christine Lagarde, IMF Managing Director Singapore Fintech Festival, November 14, 2018 Introduction Distinguished guests, ladies and gentlemen—good morning and thank you for the opportunity to participate in this important event. In Singapore, it is often windy. Winds here bring change, and opportunity. Historically, they blew ships to its port. These resupplied while waiting for the Monsoon to pass, for the seasons to change. “Change is the only constant,” wrote the ancient Greek philosopher, Heraclitus of Ephesus. Singapore knows this. You know this. It is the true spirit of the Fintech Festival—opening doors to new digital futures; hoisting sails to the winds of change. And yet change can appear daunting, destabilizing, even threatening. This is especially true for technological change, which disrupts our habits, jobs, and social interactions. The key is to harness the benefits while managing the risks. When it comes to fintech, Singapore has shown exceptional vision—think of its regulatory sandbox where new ideas can be tested. Think of its Fintech Innovation Lab, and its collaboration with major central banks on cross-border payments. In this context, I would like to do three things this morning: • First, frame the issue in terms of the changing nature of money and the fintech revolution. • Second, evaluate the role for central banks in this new financial landscape— especially in providing digital currency. • Third, look at some downsides, and consider how they can be minimized. 1. The changing nature of money and the fintech revolution Let me begin with the big issue on the table today—the changing nature of money. When commerce was local, centered around the town square, money in the form of tokens— metal coins—was sufficient. And it was efficient. 2 The exchange of coins from one hand to another settled transactions. So long as the coins were valid—determined by glancing, scratching, or even biting into them—it did not matter which hands held them. But as commerce moved to ships, like those that passed through Singapore, and covered increasingly greater distances, carrying coins became expensive, risky, and cumbersome. Chinese paper money—introduced in the 9th century—helped, but not enough. Innovation produced bills of exchange—pieces of paper allowing merchants with a bank account in their home city to draw money from a bank at their destination. The Arabs called these Sakks, the origin of our word “check” today. These checks, and the banks that went along with them, spread around the world, spearheaded by the Italian bankers and merchants of the Renaissance. Other examples are the Chinese Shansi and Indian Hundi bills. Suddenly, it mattered whom you dealt with. Was this Persian merchant the rightful owner of that bill? Was the bill trustworthy? Was that Shanxi bank going to accept it? Trust became essential—and the state became the guarantor of that trust, by offering liquidity backstops, and supervision. Why is this brief tour of history relevant? Because the fintech revolution questions the two forms of money we just discussed—coins and commercial bank deposits. And it questions the role of the state in providing money. We are at a historic turning point. You—young and bold entrepreneurs gathered here today— are not just inventing services; you are potentially reinventing history. And we are all in the process of adapting. A new wind is blowing, that of digitalization. In this new world, we meet anywhere, any time. The town square is back—virtually, on our smartphones. We exchange information, services, even emojis, instantly… peer to peer, person to person. We float through a world of information, where data is the “new gold”—despite growing concerns over privacy, and cyber-security. A world in which millennials are reinventing how our economy works, phone in hand. And this is key: money itself is changing. We expect it to become more convenient and userfriendly, perhaps even less serious-looking. We expect it to be integrated with social media, readily available for online and person-toperson use, including micro-payments. And of course, we expect it to be cheap and safe, protected against criminals and prying eyes. What role will remain for cash in this digital world? Already signs in store windows read “cash not accepted.” Not just in Scandinavia, the poster child of a cashless world. In various 3 other countries too, demand for cash is decreasing—as shown in recent IMF work. And in ten, twenty, thirty years, who will still be exchanging pieces of paper? Bank deposits too are feeling pressure from new forms of money. Think of the new specialized payment providers that offer e-money—from AliPay and WeChat in China, to PayTM in India, to M-Pesa in Kenya. These forms of money are designed with the digital economy in mind. They respond to what people demand, and what the economy requires. Even cryptocurrencies such as Bitcoin, Ethereum, and Ripple are vying for a spot in the cashless world, constantly reinventing themselves in the hope of offering more stable value, and quicker, cheaper settlement. 2. A case for Central Bank Digital Currencies Let me now turn to my second issue: the role of the state—of central banks—in this new monetary landscape. Some suggest the state should back down. Providers of e-money argue that they are less risky than banks, because they do not lend money. Instead, they hold client funds in custodian accounts, and simply settle payments within their networks. For their part, cryptocurrencies seek to anchor trust in technology. So long as they are transparent—and if you are tech savvy—you might trust their services. Still, I am not entirely convinced. Proper regulation of these entities will remain a pillar of trust. Should we go further? Beyond regulation, should the state remain an active player in the market for money? Should it fill the void left by the retreat of cash? Let me be more specific: should central banks issue a new digital form of money? A statebacked token, or perhaps an account held directly at the central bank, available to people and firms for retail payments? True, your deposits in commercial banks are already digital. But a digital currency would be a liability of the state, like cash today, not of a private firm. This is not science fiction. Various central banks around the world are seriously considering these ideas, including Canada, China, Sweden, and Uruguay. They are embracing change and new thinking—as indeed is the IMF. 4 Today, we are releasing a new paper 1 on the pros and cons of central bank digital currency— or “digital currency” for short [link, hold up copy of paper]. It focuses on domestic, not cross-border effects of digital currency. The paper is available on the IMF website. I believe we should consider the possibility to issue digital currency. There may be a role for the state to supply money to the digital economy. This currency could satisfy public policy goals, such as (i) financial inclusion, and (ii) security and consumer protection; and to provide what the private sector cannot: (iii) privacy in payments. a) Financial inclusion Let me start with financial inclusion, where digital currency offers great promise, through its ability to reach people and businesses in remote and marginalized regions. We know that banks are not exactly rushing to serve poor and rural populations. This is critical, because cash might no longer be an option here. If the majority of people adopt digital forms of money, the infrastructure for cash would degrade, leaving those in the periphery behind. What about subsidizing cash usage in those areas? But that means that economic life in the periphery would become disconnected from the center. Of course, offering a digital currency is not necessarily the only answer. There may be scope for governments to encourage private sector solutions, by providing funding, or improving infrastructure. b) Security and consumer protection The second benefit of digital currency relates to security and consumer protection. This is really a David versus Goliath argument. In the old days, coins and paper notes may have checked the dominant positions of the large, global payment firms—banks, clearinghouses, and network operators. Simply by offering a low cost and widely available alternative. Without cash, too much power could fall into the hands of a small number of outsized private payment providers. Payments, after all, naturally lean toward monopolies—the more people you serve, the cheaper and more useful the service. For a start, private firms may under-invest in security to the extent they do not measure the full cost to society of a payment failure. Resilience may also suffer—with only a few links in the payment chain, the system may stop working if one of these links breaks. Think about a cyber-attack, a glitch, bankruptcy, or a firm’s withdrawal from the local market. 1 IMF Staff Discussion Note entitled “Casting Light on Central Bank Digital Currency,” published today on our website. 5 Regulation may not be able to fully redress these downsides. A digital currency could offer advantages, as a backup means of payment. And it could boost competition by offering a low-cost and efficient alternative—as did its grandfather, the old reliable paper note. c) Privacy The third benefit of digital currency I would like to highlight lies in the privacy domain. Cash, of course, allows for anonymous payments. We reach for cash to protect our privacy for legitimate reasons: to avoid exposure to hacking and customer profiling, for instance. Consider a simple example. Imagine that people purchasing beer and frozen pizza have higher mortgage defaults than citizens purchasing organic broccoli and spring water. What can you do if you have a craving for beer and pizza but do not want your credit score to drop? Today, you pull out cash. And tomorrow? Would a privately-owned payment system push you to the broccoli aisle? Would central banks jump to the rescue and offer a fully anonymous digital currency? Certainly not. Doing so would be a bonanza for criminals. 3. Downsides of Bank Digital Currencies This brings me to my third area—the potential downsides of digital currency. The obvious ones are risks to financial integrity and financial stability. But I would also like to highlight risks of stifling innovation—the last thing you want. My main point will be that we should face these risks creatively. How might we attenuate them by designing digital currency in new and innovative ways? Technology offers a very wide canvas to do so. a) Risks to financial integrity Let’s return to the tradeoff between privacy and financial integrity. Could we find a middle ground? Central banks might design digital currency so that users’ identities would be authenticated through customer due diligence procedures and transactions recorded. But identities would not be disclosed to third parties or governments unless required by law. So when I purchase my pizza and beer, the supermarket, its bank, and marketers would not know who I am. The state might not either, at least by default. Anti-money laundering and terrorist financing controls would nevertheless run in the background. If a suspicion arose it would be possible to lift the veil of anonymity and investigate. 6 This setup would be good for users, bad for criminals, and better for the state, relative to cash. Of course, challenges remain. My goal, at this point, is to encourage exploration. b) Risks to financial stability The second risk relates to financial stability. Digital currencies could exacerbate the pressure on bank deposits we discussed earlier. If digital currencies are sufficiently similar to commercial bank deposits—because they are very safe, can be held without limit, allow for payments of any amount, perhaps even offer interest—then why hold a bank account at all? But banks are not passive bystanders. They can compete with higher interest rates and better services. What about the risk of bank runs? It exists. But consider that people run when they believe that cash withdraws are honored on a first-come-first-serve basis—the early bird gets the worm. Digital currency, instead, because it can be distributed much more easily than cash, could reassure even the person left lying on the couch! In addition, if depositors are running to foreign assets, they will also shun the digital currency. And in many countries, there are already liquid and safe assets to run toward— think of mutual funds that only hold government bonds. So, the jury is still out on whether digital currencies would really upset financial stability. c) Risks to innovation If digital currency became too popular, it might ironically stifle innovation. Where is your role if the central bank offers a full-service solution, from digital wallet, to token, to back-end settlement services? What if, instead, central banks entered a partnership with the private sector—banks and other financial institutions—and said: you interface with the customer, you store their wealth, you offer interest, advice, loans. But when it comes time to transact, we take over. This partnership could take various forms. Banks and other financial firms, including startups, could manage the digital currency. Much like banks which currently distribute cash. Or, individuals could hold regular deposits with financial firms, but transactions would ultimately get settled in digital currency between firms. Similar to what happens today, but in a split second. All nearly for free. And anytime. 7 The advantage is clear. Your payment would be immediate, safe, cheap, and potentially semi-anonymous. As you wanted. And central banks would retain a sure footing in payments. In addition, they would offer a more level playing field for competition, and a platform for innovation. Meanwhile your bank, or fellow entrepreneurs, would have ensured a friendly user experience based on the latest technologies. Putting it another way: the central bank focuses on its comparative advantage—back-end settlement—and financial institutions and start-ups are free to focus on what they do best— client interface and innovation. This is public-private partnership at its best. Conclusion Let me conclude. I have tried to evaluate the case this morning for digital currency. The case is based on new and evolving requirements for money, as well as essential public policy objectives. My message is that while the case for digital currency is not universal, we should investigate it further, seriously, carefully, and creatively. More fundamentally, the case is about change—being open to change, embracing change, shaping change. Technology will change, and so must we. Lest we remain the last leaf on a dead branch, the others having decided to fly with the wind. In the world of Fintech, we need to harness change so it is fair, safe, efficient, and dynamic. That was the goal of the Bali Fintech Agenda launched by the IMF and World Bank last October. When the winds of change pick up, what will guide us in our journey? The captains sailing through the Straits of Singapore followed the North Star. And today? Tomorrow? I suggest we follow a girl. A young girl. A fearless girl. [show picture of statue]. If you are lucky, you might be able to meet her in person in New York’s financial district. She is bold. She is brave. She is confident. She faces forward, toward the future, with grit and determination—a future she herself is going to shape, with eyes wide open, eagerly, steadily. I hear her say: Let us sail ahead. I am not afraid. (pause) I, am not afraid. Thank you. Tommaso Mancini-Griffoli, Maria Soledad Martinez Peria, Itai Agur, Anil Ari, John Kiff, Adina Popescu, and Celine Rochon With contributions from Fabio Comelli, Federico Grinberg, Ashraf Khan, and Kristel Poh DISCLAIMER: Staff Discussion Notes (SDNs) showcase policy-related analysis and research being developed by IMF staff members and are published to elicit comments and to encourage debate. The views expressed in Staff Discussion Notes are those of the author(s) and do not necessarily represent the views of the IMF, its Executive Board, or IMF management. SDN/18/08 Casting Light on Central Bank Digital Currency November 2018 IMF ST A F F D ISC U SSIO N N O T E CASTING LIGHT ON CENTRAL BANK DIGITAL CURRENCY Casting Light on Central Bank Digital Currency 1 Authorized for distribution by Tobias Adrian, Martin Muhleisen, and Maurice Obstfeld DISCLAIMER: Staff Discussion Notes (SDNs) showcase policy-related analysis and research being developed by IMF staff members and are published to elicit comments and to encourage debate. The views expressed in Staff Discussion Notes are those of the author(s) and do not necessarily represent the views of the IMF, its Executive Board, or IMF management. JEL Classification Numbers: E51, E58, E52 Keywords: money, central bank digital currencies, monetary policy Authors’ E-mail Address: TManciniGriffoli@imf.org; MMartinezPeria@imf.org; iagur@imf.org; aari@imf.org; jkiff@imf.org; apopescu@imf.org; crochon@imf.org 1 Paper prepared by an IMF Staff team from MCM, RES, SPR, and LEG, led by Tommaso Mancini-Griffoli (MCM) and Maria Soledad Martinez Peria (RES) under the guidance of Dong He (MCM), Giovanni Dell’Ariccia (RES), and Vikram Haksar (SPR). Comments are gratefully acknowledged by seminar participants at the IMF and Inter-American Development Bank, as well as by IMF reviewing divisions, and separately by Tobias Adrian, Jihad Alwazir, Tamim Bayoumi, Pelin Berkmen, Luis Brandao Marques, Jess Cheng, Chris Colford, Ulric Eriksson von Allmen, Gaston Gelos, Masaru Itatani, Nigel Jenkinson, Tanai Khiaonarong, Darryl King, Amina Lahreche, Ross Leckow, Rodolfo Maino, Fabiana Melo, Aditya Narain, Maurice Obstfeld, Luca Ricci, James Roaf, Herve Tourpe, Romain Veyrune, and Froukelien Wendt. Karen Lee (MCM) provided excellent research assistance. We are grateful to Shanti Karunaratne, Danica Owczar, and Wifianni Wirsatyo (MCM) and Gabriela Maciel (RES) for outstanding administrative assistance. We are indebted to Joe Procopio (COM) for copyediting the document. 2 INTERNATIONAL MONETARY FUND CASTING LIGHT ON CENTRAL BANK DIGITAL CURRENCY CONTENTS EXECUTIVE SUMMARY __________________________________________________________________________ 4 I. INTRODUCTION _______________________________________________________________________________ 6 II. BASICS OF CENTRAL BANK DIGITAL CURRENCY ___________________________________________ 7 III. A CONCEPTUAL FRAMEWORK TO COMPARE DIFFERENT FORMS OF MONEY ___________ 9 IV. IS THERE A ROLE FOR CBDC? USER PERSPECTIVE_________________________________________ 14 V. IS THERE A ROLE FOR CBDC? CENTRAL BANK PERSPECTIVE______________________________ 15 Social Criteria for Money ________________________________________________________ 15 Would CBDC Undermine Financial Integrity? ____________________________________ 20 Would CBDC Undermine Financial Stability and Banking Intermediation? _______ 21 Would Monetary Policy Transmission Remain Effective? ________________________ 25 VI. CENTRAL BANK RESEARCH AND EXPERIMENTS __________________________________________ 27 VII. CONCLUSION ______________________________________________________________________________ 30 REFERENCES ____________________________________________________________________________________ 32 APPENDIX ______________________________________________________________________________________ 38 INTERNATIONAL MONETARY FUND 3 CASTING LIGHT ON CENTRAL BANK DIGITAL CURRENCY EXECUTIVE SUMMARY Digitalization is reshaping economic activity, shrinking the role of cash, and spurring new digital forms of money. Central banks have been pondering whether and how to adapt. One possibility is central bank digital currency (CBDC)—a widely accessible digital form of fiat money that could be legal tender. While several central banks have studied the adoption of CBDC and have undertaken pilots, many have not actively explored it and remain skeptical. This discussion note proposes a conceptual framework to assess the case for CBDC adoption from the perspective of users and central banks. It abstracts from cross-border considerations by assuming that CBDC is for domestic use only. This note discusses possible CBDC designs, and explores potential benefits and costs, with a focus on the impact on monetary policy, financial stability, and integrity. This note also surveys research and pilot studies on CBDC by central banks around the world. The main takeaways are as follows: • The impact of CBDC introduction will hinge on its design and country-specific characteristics. Critical features will be anonymity (the traceability of transactions), security, transaction limits, and interest earned. The role of cash and commercial bank deposits in payments will also matter. • CBDC could strengthen the benefits and reduce some of the costs and risks to the payment system and could help encourage financial inclusion. However, demand will not necessarily be very high and will depend on the attractiveness of alternative forms of money. Moreover, there are other payment solutions to help central banks more fully achieve their goals relative to money. CBDC will have to contend with operational risks arising from disruptions and cyberattacks. • Token-based CBDC—with payments that involve the transfer of an object (namely, a digital token)—could extend some of the attributes of cash to the digital world. CBDC could provide varying degrees of anonymity and immediate settlement. It could thus curtail the development of private forms of anonymous payment but could increase risks to financial integrity. Design features such as size limits on payments in, and holdings of, CBDC would reduce but not eliminate these concerns. • Account-based CBDC—with payments through the transfer of claims recorded on an account— could increase risks to financial intermediation. It would raise funding costs for deposit-taking institutions and facilitate bank runs during periods of distress. Again, careful design and accompanying policies should reduce, but not eliminate, these risks. • CBDC is unlikely to affect monetary policy transmission significantly, although operations may need adaptation. Transmission could strengthen if CBDC spurs greater financial inclusion. Interest-bearing CBDC would eliminate the effective lower bound on interest rate policy, but only with constraints on the use of cash. 4 INTERNATIONAL MONETARY FUND CASTING LIGHT ON CENTRAL BANK DIGITAL CURRENCY • Overall, it is too early to draw firm conclusions on the net benefits of CBDC. Central banks should consider their specific country circumstances, paying careful attention to the risks and relative merits of alternative solutions. Further analysis of technological feasibility and operational costs is needed. INTERNATIONAL MONETARY FUND 5 CASTING LIGHT ON CENTRAL BANK DIGITAL CURRENCY I. INTRODUCTION 1. The impact of digitalization is widespread and profound. It is changing the nature of jobs, education, commerce, innovation, and product life cycles. Demographics are accelerating these developments. Millennials now outnumber baby boomers (Tilford 2018) and are steering the economy toward their world—one in which digital platforms are central, and nearly second nature. 2. Payments, and more fundamentally money, are also undergoing tremendous change. 2 Technology, new employment arrangements, and the growing decentralized service economy, as well as evolving social attitudes, are driving efforts to build new and more decentralized forms of money. These offer peer-to-peer transactions, micropayments, and easy-to-use interfaces integrated with social networks. Payments are increasingly being diverted toward privately run solutions. Even cryptocurrencies such as Bitcoin, Ethereum, and Ripple—still early in their development cycle—offer competing forms of money. 3. Deep and pressing questions arise. Is there a role for cash, or a cash-like form of money, in the digital world? Should central banks offer new forms of money? If so, what are the implications for monetary policy and financial intermediation, stability, and integrity? 4. Central banks are taking these questions seriously. Several are actively investigating the possibility of a central bank digital currency (CBDC). This new central bank liability would be a widely accessible digital form of fiat money, intended as legal tender. One day, it could fully replace physical cash. CBDC seems to be a natural next step in the evolution of official coinage (from metalbased money, to metal-backed banknotes, to physical fiat money). 5. This note offers a conceptual framework to evaluate the case for CBDC. The approach aims to answer a simple question: Does CBDC offer benefits? On the demand side, would it satisfy end user needs better than other forms of money? And on the supply side, would issuing CBDC allow central banks to more effectively satisfy public policy goals, including financial inclusion, operational efficiency, financial stability, monetary policy effectiveness, and financial integrity? In short, is CBDC a desirable form of money given existing and rapidly evolving alternatives? Although its adoption appears more promising in some circumstances than in others, a final decision requires careful evaluation of country-specific circumstances, including a review of technological feasibility and costs. These factors are beyond the scope of this note. 6. This note includes a summary of pilot projects and analyses from central banks exploring the possibility of issuing CBDC. The analysis is based on publicly issued materials and discussions with staff members at central banks and technology providers around the world. 7. This note contributes to a growing body of literature on CBDC. Others have already explored this topic, including international organizations such as the Bank for International Settlements3 and several central banks (Bank of Canada, People’s Bank of China, Bank of England, 2 See He and others (2017) for a discussion of the impact of financial technology on financial services. 3 See the report produced by the Bank for International Settlements Committee on Payments and Market Infrastructures (CPMI 2018). 6 INTERNATIONAL MONETARY FUND CASTING LIGHT ON CENTRAL BANK DIGITAL CURRENCY Bank of Finland, Norges Bank, Danmarks Nationalbank, Sveriges Riksbank), 4 as well as academics and policymakers (Raskin and Yermack 2016; Rogoff 2016; Bordo and Levin 2018; He 2018; He and Khan 2018; Kahn, Rivadeneyra, and Wong 2018). Those analyses discuss the pros and cons of CBDC adoption and, in some cases, make policy recommendations. Relative to those studies, this note makes several contributions. First, it introduces a conceptual framework to evaluate the desirability of CBDC. In doing so, it emphasizes the perspective of users and their preferences for different features of money, in addition to the goals of central banks. Second, it considers policy responses to CBDC adoption to mitigate ensuing risks. Third, it offers a one-stop reference on views and ongoing and future plans of some central banks from around the world regarding CBDC. 8. The remainder of the note is organized in six sections. Section II covers the basics of CDBC. Section III lays out a conceptual framework to compare different forms of money from the standpoint of end users and the central bank. Section IV puts the approach to use, evaluating whether there might be a role for CBDC from the perspective of users. Section V considers that question from the viewpoint of central banks, and in so doing explores different options for designing CBDC. Part of this section is devoted to gauging the impact of CBDC on financial integrity, financial stability, and monetary policy transmission. Section VI offers an overview of central bank investigations, and the last section concludes and raises questions for future research on how CBDC might affect cross-border payments. II. BASICS OF CENTRAL BANK DIGITAL CURRENCY 9. CBDC is a new form of money, issued digitally by the central bank and intended to serve as legal tender. 5 It would differ, however, from other forms of money typically issued by central banks: cash and reserve balances. CBDC designed for retail payments would be widely available. In contrast reserves are available only to selected institutions, mostly banks with accounts at the central bank. 6 Clearly, CBDC is not intended to have a physical form like cash. But as cash, it would be widely accessible to a country’s residents—and potentially to individuals and organizations abroad. 7 CBDC could be used as easily for person-to-person, person-to-business, and business-tobusiness transactions of any amount, a notable improvement over cash. 8 Reserves, by contrast, settle wholesale interbank payments only. 4 Fung and Halaburda (2016); Bech and Garratt (2017); Engert and Fung (2017); Davoodalhosseini (2018); Yifei (2018); Carney (2018); Kumhof and Noone (2018); Grym and others (2017); Norges Bank (2018); Gürtler and others (2017); and Sveriges Riksbank (2017). 5 Change in legislation may be needed for CBDC to be legal tender. The definition of legal tender—usually applied to banknotes and coins issued by central banks—varies slightly across jurisdictions. For instance, a creditor is not obligated to accept payment in legal tender in all jurisdictions. See He and others (2016) for details. 6 Reserves can be interpreted as a wholesale form of CBDC used exclusively for interbank payments. 7 The cross-border implications of CBDC are not considered here, but some important questions are put forth in the concluding section for future work. 8 Person to person transactions are deliberately defined as person to person and not peer to peer. The first suggests that payments can be made seamlessly between individuals, such as when splitting a dinner bill. The second, INTERNATIONAL MONETARY FUND 7 CASTING LIGHT ON CENTRAL BANK DIGITAL CURRENCY 10. CBDC could be account- or token-based, the former involving the transfer of a claim on an account and the latter of a token between wallets. 9 A transaction in account-based CBDC would resemble today’s transactions between commercial bank depositors, except accounts would be held with the central bank. A payer would log in to an account at the central bank—for example, through a web page or an app on a handheld device—and request a transfer of funds to a recipient’s account, also at the central bank (Figure 1). The central bank would ensure settlement by updating a master ledger, but only after verification of the payer’s authority to use the account, sufficient funds, and authenticity of the payee’s account. The exchange of information would therefore be substantial. Figure 1. Account- and Token-Based CBDC, Basic Mechanics Source: IMF staff 11. Transacting in token-based CBDC would involve more steps than exchanging cash but would offer the convenience of not having to meet in person. Unlike cash—the prime example of a traditional token-based form of money—CBDC tokens would be too complex to be distinguished from counterfeits by parties to the transaction. Settling a transaction using tokenbased CBDC would require external verification of the tokens. As a result, transactions might not be entirely anonymous, like cash. The extent of anonymity would depend on whether wallets are registered and transaction information is recorded. however, is often used to denote payments in cryptocurrency for which transaction parties, or “peers,” are also involved in settlement. 9 The distinction between accounts (intangible property) and tokens (tangible property) is emphasized in Kahn and Roberds (2009). See also Kahn, Rivadeneyra, and Wong (2018) for a discussion of different forms of token-based CBDC. 8 INTERNATIONAL MONETARY FUND CASTING LIGHT ON CENTRAL BANK DIGITAL CURRENCY 12. Verification of the tokens and settlement could be centralized or decentralized, depending on the technology used. Decentralized settlement is possible via the use of distributed ledger technology (DLT). However, although the technology is evolving, it currently falls short in scalability, energy efficiency, and payment finality (He and others 2017). DLT could be used over a closed (“permissioned”) network managed by the central bank. But there are other types of centralized settlement technology that may prove more efficient. These would check the validity of the tokens’ serial numbers, then reassign numbers once tokens change wallets to avoid the risk of double spending. III. A CONCEPTUAL FRAMEWORK TO COMPARE DIFFERENT FORMS OF MONEY 13. The conceptual framework in this discussion note is geared toward tackling one key question: Does CBDC offer net benefits? We explore two sides to benefits. First, on the demand side, to what extent might CBDC satisfy end users’ need for money, and would it do so better than other forms of money? Second, on the supply side, would CBDC allow central banks to more fully achieve their public policy goals and overcome specific market failures? In answering these questions and to assess the net benefits of CBDC, we also explore some potential costs and tradeoffs associated with its adoption. We refrain, however, from assessing technical implementation costs and feasibility. 14. The case for CBDC hinges on whether it strengthens the functions of money for users and central banks. In economics, money is seen as having three functions: a unit of account, a means of payment, and a store of value. As a unit of account, money serves as a measuring stick ideally linked to the same basket of goods over time. As a means of payment, it facilitates transactions. As a secure store of value, it provides refuge from various sources of risk defined further below. 10 15. Users will seek a form of money that maximizes private benefits and minimizes associated costs and risks. 11 Related criteria are listed below (Figure 2). Emphasis is placed on the means of payment and store of value functions of money, for which criteria are more diverse. The relative weight of each criterion below will vary by country and user. 10 These three properties are not entirely independent. If money is not a store of value, it will certainly not offer a satisfactory means of payment and will consequently not be a satisfactory unit of account. Money offers substantial efficiency gains by helping coordinate the specification of financial contracts. See He and others (2016) for a discussion of money versus “currency,” a term used for money issued by governments as legal tender. 11 Kahn, Rivadeneyra, and Wong (2018) adopt a similar approach—“convenience, costs, and safety”—in their phrasing. INTERNATIONAL MONETARY FUND 9 CASTING LIGHT ON CENTRAL BANK DIGITAL CURRENCY Figure 2. User Criteria to Judge Different Forms of Money Maximize benefits Means of payment Store of value • Liquidity: Payment on demand • Returns: Nominal interest payments 12 • Scalability: Payment of any size (no limits) • Acceptance: Person to person, person to business, business to business to and from any device; no network limitation • Extra services: Preferential access to other financial services (loans, advice, etc.) Minimize costs • Transaction: Ease of use; fees Minimize risks • Settlement: Lag between agreeing to a transaction and actual receipt of funds • Disclosure: Degree of anonymity • Theft: Ability to reverse fraudulent transactions, exposure to fraud/cyber risk • Loss: Ability to claim ownership or recover access if lost • Default: of the money issuer 16. One important criterion stands out: the ability to make anonymous transactions. Regarding money, anonymity covers the extent to which identity and transactions are, or can be, disclosed to transaction parties, third parties, and the government. There are legitimate reasons people may prefer at least some degree of anonymity—potentially when it comes to everyone except the government, and regarding the government unless a court order unlocks encrypted transaction information. It is a way to avoid customer profiling—commercial use of personal information, for example, to charge higher mortgage rates to people who purchase alcohol. Another advantage of anonymity is limiting exposure to hacking. Moreover, anonymity is often associated with privacy—widely recognized as a human right (as stated in the Universal Declaration of Human Rights [Article 12] and elsewhere). 17. On the supply side, central banks play a pivotal role and ensure that money delivers on its three functions. For central banks, this role means two things. First, because they are accountable to the public, central banks must design the money they issue—and regulate private forms of money—in a way that satisfies the user needs stated earlier. Second, because they are public policy institutions, they must ensure that money also meets important social criteria (illustrated in Figure 3 and discussed in Box 1): • 12 As a unit of account, money is an important public good that requires price stability in all economic circumstances. The design of money can favor or interfere with this goal. For instance, because cash pays no interest, central banks find it difficult to offer deeply negative interest rates following sharp recessions (more on this later). Bordo and Levin (2018) argue that for money to be a secure store of value, it must offer risk-free returns. 10 INTERNATIONAL MONETARY FUND CASTING LIGHT ON CENTRAL BANK DIGITAL CURRENCY • As a means of payment, money must be universally available and verifiable as well as efficient, while ensuring appropriate consumer protection and minimal cost to taxpayers. • As a store of value, money must be as secure as possible, but it must also allow for efficient allocation of resources. 18. In addition, central banks will prefer forms of money that support, or at least do not undermine, three other public policy goals: financial integrity, financial stability, and monetary policy effectiveness. 13 In turn, each of these further supports the three functions of money. Financial integrity covers, among other things, anti–money laundering and combating the financing of terrorism (AML/CFT) rules, including customer due diligence measures and additional measures aimed at fighting corruption and fostering good governance. Figure 3. The Central Bank’s Criteria to Evaluate Different Forms of Money Box 1. The Birth of Central Banking—a Quest for Efficient, Low-Cost, and Safe Money There is a fascinating historical account of the birth of central banking in Europe in Kahn, Quinn, and Roberds (2014). Coins were the predominant form of money in medieval times and during the early Renaissance. Although they were a relatively efficient means of payment and allowed immediate settlement, transportation costs grew substantially as commerce spread geographically. Bills of exchange were introduced to decrease these costs; merchants could purchase bills drawn on the foreign banks where they intended to do business. However, bills carried substantial counterparty risk, as debtor banks often did not pay. The Bank of Amsterdam was established in 1609 to reduce the risks inherent in bills of 13 Other goals of central banks regarding CBDC include measures to stem the loss of seigniorage from the growth of new forms of private money. CBDC would in fact preserve seigniorage and possibly increase it if the central bank’s balance sheet grew, depending on whether CBDC earned interest. However, because seigniorage is small in many countries, we do not entertain this line of argument further. INTERNATIONAL MONETARY FUND 11 CASTING LIGHT ON CENTRAL BANK DIGITAL CURRENCY exchange, following earlier (though smaller-scale and private) models, such as the Taula de Canvi in Barcelona (1401), Banco di San Giorgio in Genoa (1408), and Banco di Rialto in Venice (1587). The Bank of Amsterdam stood ready to pay high-quality Dutch guilders against the receipt of bills. In 1683, it offered account balances to settle bills of exchange on its books, and soon thereafter suspended the right of coin withdrawals entirely, thereby shifting the unit of account from commodity money to its liabilities, as in a system with metal-backed banknotes. At its peak in the mid-1700s, yearly turnover through the accounts at the Bank of Amsterdam was more than 2.5 times Dutch GDP. This evolution highlights ongoing efforts to establish increasingly efficient, secure, and lower-cost forms of money. The Bank of Amsterdam first reduced the risks and costs of payments by offering immediate and final settlement. Then it did more by offering liquidity backstops to merchant banks that were funded with bills of exchange. Without this backstop, merchant banks would have faced a trade-off. They either had to offer a secure store of value by storing the cash received for their bills of exchange or efficiently allocate funds by lending the cash. This would expose them and their creditors to the liquidity risk associated with maturity mismatches. The Bank of Amsterdam’s innovation underscores the role of the state in enhancing the safety of private payment systems without contravening their efficiency. Bank supervision, lender of last resort functions, and deposit insurance do much the same today. 19. The last leg of the conceptual framework is to determine competitors to CBDC. These fall into four categories, but will vary by country: cash, commercial bank deposits, narrow finance, and cryptocurrencies. All except cash are evolving and rapidly gaining market share. 20. Commercial bank deposits are going through notable improvements. Payments have traditionally been facilitated by debit card networks. Today, two continuing transformations are notable, especially in advanced economies. The first stems from “wrapper” technology, such as Venmo in the United States, which allows transactions to take place between mobile devices (bypassing expensive point-of-sale terminals) and adds a layer of security. 14 The other is centralbank-provided fast-payment solutions (“fast payments”). 15 These allow payments of any size and type (person to person, person to business, business to business) to be settled instantaneously by the central bank in reserve money through a dedicated platform running continuously at negligible cost. 16 14 Touchless technology also facilitates debit card transactions. There is wrapper technology, such as PayPal, for credit card transactions as well, but credit-based transactions lie outside the scope of this note. 15 The largest-scale project is the European Central Bank’s Target Instant Payment Settlement service, to be introduced in November 2018. There are already other similar initiatives, such as Hong Kong SAR’s Faster Payments introduced in August 2018, Sweden’s Swish, Denmark’s Straksclearing, and Australia’s New Payments Platform. Other so-called fast or instant payment solutions are also being rolled out exclusively by commercial banks (see CPMI 2016). 16 Fast payments can be thought of as a form of CBDC offered through a public-private partnership, because they allow people to settle in central bank reserves at will, at any time, through banks (resembling the proposal in Bordo and Levin 2018). The central bank, then, offers the means of payment function of money and banks the store of value function. Together, they offer money’s three functions. From a technological standpoint, fast payments would however differ, even if account-based. Fast-payment engines are optimized for interoperability with real-time gross 12 INTERNATIONAL MONETARY FUND CASTING LIGHT ON CENTRAL BANK DIGITAL CURRENCY 21. “Narrow finance solutions” is a term introduced in this discussion note to capture the various new forms of private money backed one for one by central bank liabilities, either cash or reserves. These offer stable nominal value, security, liquidity, and potentially close to a risk-free rate of return. The parallel here is with currency boards (such as in Hong Kong SAR) or metal-backed banknote systems (such as the gold standard). Two versions of narrow finance solutions are relevant. The first is stored value facilities 17 such as AliPay and WePay in China, PayTM in India, M-Pesa in Kenya, and Bitt.com in the Caribbean. These provide private e-money to users against funds received and placed in custodian accounts. Transactions occur between electronic wallets installed on handheld devices, can be of any size (although they are usually not large), and are centrally cleared, but are restricted to participants in the same network. However, holding these forms of money entails some risk. 18 Nonetheless, this segment is gaining widespread and very rapid acceptance. The second version of narrow finance solutions—narrow banks—is only beginning to materialize. It covers institutions that invest client funds only in highly liquid and safe government assets—such as excess reserves at the central bank—and do not lend. However, they allow payments in their liabilities through debit cards or privately issued digital money. 19 22. Cryptocurrencies are different along many dimensions and struggle to fully satisfy the functions of money, in part because of erratic valuations. Examples include Bitcoin, Ethereum, and Ripple. These currencies are not the liability of any institution and are not backed by assets. Their value is usually volatile, because most have rigid issuance rules. Some new cryptocurrencies attempt to stabilize their value by controlling issuance according to a function of price deviations from a fiat currency or commodity (as in an exchange rate peg). Examples are Basecoin and Stablecoin. 20 In all cases, transactions are settled in a decentralized fashion, using distributed ledger technology. settlement systems (in which banks keep liquidity and borrow funds to settle intraday payments) and, by extension, with private banks. This integration would not be necessary for account-based CBDC. 17 The terms “stored value facilities” and “e-money” are taken from the Monetary Authority of Singapore (2017); emoney is defined as “electronically stored monetary value in a payment account that can be used to purchase goods or services, or to transfer funds to another individual.” 18 These depend on whether the stored value facility has access to the funds in the custodian account and whether it can invest them in illiquid assets and the degree to which e-money issuance can exceed reserves. Both of these examples invite the question of why users would choose to exchange or forgo a safe and liquid asset like cash or a government bond for another provided by a private intermediary. The answer lies in the ease with which users can initiate payments on private platforms, and potentially access other services. Note that new players—such as large tech firms, including Amazon, Apple, Google, and Facebook—may well enter this space. They could offer e-money to purchase their goods at a discount, in exchange for valuable information. 19 See Ali (2018) for instance for a solution based on privately issued digital money. Note that firms in the narrow finance category do not create money, unlike fractional reserve banks. In a world of CBDC alone, only the central bank could create money. Today most money creation is “outsourced” to commercial banks, which create deposits when they extend credit. The process of money creation nonetheless responds to interest rates set by central banks. 20 See, however, Eichengreen (2018) for doubts about the ability to maintain a peg, short of full backing by fiat currency as in the narrow finance example (referred to as a currency board). INTERNATIONAL MONETARY FUND 13 CASTING LIGHT ON CENTRAL BANK DIGITAL CURRENCY IV. IS THERE A ROLE FOR CBDC? USER PERSPECTIVE 23. Competing forms of money can be ranked according to the criteria described earlier. The evaluation is presented visually in Figure A1.1 of Appendix I, in the form of “spider charts,” and explanations are provided in the subsequent Table A1.1. Higher scores are captured by points farther from the center. The more attractive the form of money, the larger the surface covered by the spider chart. All monies were deemed liquid, so that criterion was left aside. • Cash is not an especially attractive means of payment given its high transaction costs (the need to meet in person and withdraw cash, which may be difficult in remote locations), vulnerability to theft, and lack of returns in the form of interest. However, cash offers immediate settlement, no default or cyber risk, and—importantly—full anonymity, a potentially attractive feature to users. • Cryptocurrencies are the least attractive option, receiving a low score in settlement speed because of current technological limitations (which may eventually be overcome). Their main advantage is anonymity. • Private e-money provided by stored value facilities scores high on several fronts. It offers widespread acceptance, low transaction costs via user-friendly interfaces designed by customercentric firms, and full-service bundling with other financial as well as social services. Narrow banks could further reduce default risk—and possibly enhance scalability by offering largervalue payments, although potentially lower ease of use, depending on design. • Commercial bank deposits, as they were traditionally structured, offered average value. On the positive side, they provided security from theft and loss, and integration with additional services, while on the negative side there was limited acceptance (cumbersome peer-to-peer payments requiring checks or wire transfers) and scalability (floors on debit card payments). However, recent reforms and innovations, including fast payments, have improved the attractiveness of commercial bank deposits considerably. Deposits with fast payment are completely scalable and are widely accepted, without network limits, have no settlement risk, and offer limited default risk where deposit insurance is available. 21 Transaction costs (fees and ease of use) are minimal. 24. CBDC would not strictly dominate any of these alternative forms of money. As evident in Figure A1.1, CBDC would closely compete with evolving commercial bank deposits and e-money. CBDC stands out only when it comes to anonymity and default risk. There are two sets of features: fixed features, for which the central bank does not have discretion, and flexible features. • Fixed design features: CBDC would be on par with fast-payment solutions regarding acceptance (person to person, person to business, business to business without network restrictions), settlement risk, and transaction cost (to the extent that mobile interfaces are well designed—no small feat). CBDC would probably be superior regarding default risk, although in many jurisdictions only marginally. Instead, CBDC would score poorly in terms of offering additional 21 Insurance limits nonetheless should ensure that funds earmarked for payments—not all those held as savings—are protected. Households—large firms to a lesser degree—have the option of splitting deposits across accounts. 14 INTERNATIONAL MONETARY FUND CASTING LIGHT ON CENTRAL BANK DIGITAL CURRENCY services (although banks could provide the front-end applications to manage CBDC and could cross-sell services). • Flexible design features: CBDC could offer competitive interest returns, protection from theft and loss, scalability, and anonymity, though not all at once as these features depend on one another. As discussed earlier, anonymity could be provided to different degrees: relative to transaction parties, third parties, and/or the government. But the greater the anonymity, the harder it is to reverse fraudulent transactions (risk of theft) and claim ownership (risk of loss). And central banks may be prepared to offer full anonymity only with strict and low limits on CBDC holdings, thus undermining scalability. Finally, CBDC could offer interest. If it paid the policy rate, it would be as attractive as narrow finance solutions, although commercial banks could still offer higher rates on deposits (and recoup profits by charging higher lending rates—more on this later). 25. In summary, demand for CBDC will depend on its design. It may not be high in more advanced economies, except as a cash replacement, but could be very attractive elsewhere. Commercial bank deposits with fast payments, and to some extent narrow banking solutions, will rival CBDC and could be superior in some areas. CBDC could excel only regarding anonymity, although at some cost to scalability and security. As such, it would compete mostly with cash, by allowing small-value transactions with at least some degree of anonymity. Someday CBDC may, indeed, be introduced as a replacement for paper bills, which will become increasingly anachronistic as economic activity grows ever more digital. However, in jurisdictions with limited banking penetration and unreliable settlement platforms, CBDC may be more attractive to users, especially in the absence of stored value facilities. V. IS THERE A ROLE FOR CBDC? CENTRAL BANK PERSPECTIVE 26. This section gauges whether central banks could benefit from CBDC to more fully achieve public policy goals. These include satisfying the social dimensions of money’s three functions, as well as financial integrity, financial stability, and monetary policy effectiveness. Social Criteria for Money 27. CBDC is unlikely to offer near-term assurance of price stability in all economic circumstances—as needed to bolster the unit of account function of money. The global financial crisis starkly illustrated that interest rate policy can be constrained by the presence of cash. Policy rates cannot be brought significantly below zero without risking a massive shift into cash, which ensures zero returns (see Dell’Ariccia and others (2017) for a full discussion and Habermeier and others (2013) for an overview of unconventional monetary policy). However, cash is unlikely to go away anytime soon, for political reasons. Only over the longer term, if CBDC is adopted widely, would cash be eliminated, much as happened with metal coinage. In that scenario, policy rates could go deep into negative territory. But this would happen only to the extent that CBDC also charged negative rates and did not replace cash as a means to circumvent policy. In the interim, other INTERNATIONAL MONETARY FUND 15 CASTING LIGHT ON CENTRAL BANK DIGITAL CURRENCY measures have been proposed to allow for deeply negative policy rates, but with questionable feasibility and without necessarily requiring CBDC. 22 28. CBDC could help ensure equal access to a means of payment for all citizens, thereby favoring financial inclusion. But other solutions also exist and may be more efficient. Much depends on the barriers to financial inclusion. For one, cash may be difficult to obtain and use in underpopulated and rural areas. As a majority of the population shifts to digital forms of money, the infrastructure for cash (such as distribution networks, counting machines, and armored services) may deteriorate, and businesses may resist dealing with it. As in other cases of negative externalities, government intervention may be warranted. One approach may involve subsidizing the provision of cash in underserved areas. CBDC may not be a viable solution if access to technology is limited. If the problem is instead the supply of bank accounts—which banks deem unprofitable or require unaffordable or nonexistent technology—the question is whether the private sector can offer alternative solutions. M-Pesa in Kenya and PayTM in India are examples of successful initiatives, although with some state support in the case of M-Pesa (from the UK Department for International Development). Alternatively, the government could subsidize the deployment of bank branches. Or—short of direct intervention—it could facilitate the development of online banking and communication infrastructure in rural areas and reduce the cost of bank-intermediated small-value transactions by deploying fast payments. Where these solutions are not feasible, CBDC could provide an alternative. If, however, barriers to financial inclusion stem from an aversion to formalization, neither CBDC nor other initiatives would prove satisfactory. 29. CBDC may instead help reduce costs associated with the provision of cash, thereby ensuring an efficient means of payment from a public policy perspective. Issuing and managing cash are expensive. Hasan, De Renzis, Schmiedel (2013) estimate the cost to be 0.5 percent of GDP for the euro area, similar to the cost in Canada (Kosse and others 2017) and Uruguay (Alvez, Lluberas, and Ponce 2018). Costs fall mostly on banks, firms, and households. Although introducing and maintaining CBDC would probably entail substantial fixed costs, marginal operational costs would likely be low, despite the need for customer service. On this basis alone, the business case to adopt CBDC would probably be better for larger jurisdictions able to absorb the fixed costs. 30. CBDC would not help resolve the tension central banks face between offering a secure store of value and promoting financial intermediation. Narrow finance solutions offer a liquid and secure store of value at the cost of financial intermediation. This is because payments must be entirely prefunded, as explained earlier. Fractional reserve banks, however, pool the liquidity buffers households and firms maintain to respond to payment shocks. And because not all shocks materialize at once (at least not most of the time), they can lend a portion of the funds, keeping only 22 Cash could be prohibited altogether as argued in Rogoff (2014), made costly to hold as suggested in Bordo and Levin (2018), or made to depreciate against CBDC, which would become the sole legal tender, as in Agarwal and Kimball (2015). Note that if CBDC were not interest bearing, the effective lower bound could bind at even higher rates of interest, as CBDC could be stored more cheaply than cash. CBDC has also been touted as a means to implement aggressive monetary stimulus known as a “helicopter drop” by crediting CBDC accounts or wallets holding CBDC tokens. However, doing so would not necessarily reach all citizens. Moreover, the issue of legitimacy remains: how does the central bank decide how much to transfer to each household given the notable and very explicit redistributional consequences? Finally, helicopter drops would continue to be viewed as a form of monetary financing, thus undermining central bank independence. 16 INTERNATIONAL MONETARY FUND CASTING LIGHT ON CENTRAL BANK DIGITAL CURRENCY a fraction in highly liquid and safe assets. Even though central banks should be concerned if narrow banking solutions grow substantially, CBDC would not help reverse the trend. CBDC is also, after all, a form of money that requires full prefunding. Instead, fast payments would help fractional reserve banks offer money that competes with that offered by stored value facilities and other narrow finance solutions. 31. Other potential benefits of CBDC must be seen against the backdrop of a reduction in the use of cash. The trend is already evident in some countries and is expected to become more widespread. Sweden is probably the most striking example. 23 In other countries, cash in circulation as a share of GDP has actually increased in the past decade, as documented in Bech and others (2018). However, a second look reveals that such movements are largely cyclical and can be partly explained by low interest rates. As shown in Box 2, the preference for cash has mostly been decreasing or has remained unchanged except in reserve currency countries (Switzerland and the United States). In addition, the demand for cash is likely to diminish as older generations give way to more technology-adept generations. Merchants and banks in both advanced and developing economies are also trying to discourage cash transactions, given the related costs. 23 Cash in circulation as a share of GDP is currently half its value 10 years ago, representing merely 6 percent of central bank liabilities and 2 percent of the money supply. INTERNATIONAL MONETARY FUND 17 CASTING LIGHT ON CENTRAL BANK DIGITAL CURRENCY Box 2. A Closer Look at the Demand for Cash In many countries, currency in circulation has increased significantly in the past decade (Figure 4a). 24 However, much of that increase seems cyclical, because it can be explained in good part by lower interest rates, higher uncertainty, and economic recoveries. 25 Figure 4b shows that in selected countries the unexplained (residual) component of cash in circulation—loosely associated with preferences—has often decreased or remained stable. Only in Switzerland and the United States has this component increased markedly over the past decade, in part thanks to those countries’ reserve currency status. However, cash still accounts for a large share of transaction volumes, though not of value, even in advanced economies. The use of cash is subject to habit, increases with age, and decreases with education and income. It is generally lowest in countries with the most developed payment systems. 26 Figure 4a. Cash in Circulation (as percent of GDP) Source: IFS, Haver 24 See also Bech and others (2018) and Jobst and Stix (2017). 25 The baseline is a regression of the log of real cash in circulation on deposit rates and the log of real GDP. Results are robust to inclusion of the log of stock prices (to capture wealth effects), different measures of uncertainty (Chicago Board Options Exchange Volatility Index and uncertainty indices), and country-by-country and panel estimates. 26 For related analysis, see Flannigan and Parsons (2018); Arango-Arango and others (2018); Esselink and Hernández (2017); Wakamori and Welte (2017); Bagnall, Bounie, and Huynh (2016); Krüger (2016); Bennet, Schuh, and Schwartz (2014); and Sisak (2011). For example, the cash share of payments is 82 percent by volume in Austria (63 percent by value), but 46 percent in the United States (23 percent by value). 18 INTERNATIONAL MONETARY FUND CASTING LIGHT ON CENTRAL BANK DIGITAL CURRENCY Figure 4b. Growth of Real Cash Balances (in percent, annualized growth rates) Sources: Staff calculations 32. As the use of cash wanes, there is greater concern about the security of the payment system if it is dominated by private firms. But there are solutions other than CBDC. Private issuers of money, including rapidly growing stored value facilities, may not internalize the social cost of systemic disruptions from cyberattacks or negligence, and thus may not adequately invest in security (to the extent that users remain oblivious to these risks and do not require sufficient safeguards). Central banks would be more prone to do so, but that does not mean that the risk of cyberattacks would disappear. Moreover, CBDC could offer a backup solution if disruptions in the digital infrastructure materialize, unless there are large-scale natural disasters and power outages. (Cash is similarly at risk since its infrastructure also requires electricity.) Another solution is to deploy fast payments, which also give central banks control over an essential piece of the payment architecture. Finally, adequate regulation should also bolster the security of money, even if privately provided. 33. The disappearance of cash could also pose risks to consumer protection. Again, CBDC can help, but so would fast payments and regulation. Modern payment systems are often operated by a few commercial banks and by even fewer clearinghouses and messaging services. Payment systems tend to become natural monopolies. This reflects strong network externalities (the value of using a given payment network is greater the larger the user community), decreasing average costs (savings from netting transactions over a large user base), high fixed development and maintenance costs, and significant gains from aggregating data, which—to an individual—is INTERNATIONAL MONETARY FUND 19 CASTING LIGHT ON CENTRAL BANK DIGITAL CURRENCY worth little. However, private monopolistic providers will tend to offer inadequate and expensive services and could take unfair advantage of data. The prevalence of cash as an attractive and lowcost competitor may have limited the monopoly power of private monies. In the future, if antitrust regulation and data protection prove insufficient, CBDC could serve that same purpose. However, so too could the deployment of low-cost fast payments. 34. In summary, while central banks could benefit from CBDC to more fully satisfy some of the social criteria of money, in many countries there are also other solutions. CBDC may be a way to reduce the cost to society associated with the use of cash. Financial inclusion may also benefit, if private sector solutions and policy efforts do not bear fruit. However, CBDC will not support efficient allocation of resources. It could, under some circumstances, help central banks bolster the security of payment systems and consumer protection. But regulation and, where possible, fast-payment platforms should offer compelling alternatives. Can CBDC Balance Privacy and Financial Integrity Concerns? 35. In designing money, national authorities already face a trade-off between satisfying legitimate user preferences for privacy and mitigating risks to financial integrity. Cash protects privacy because it is anonymous: no account is necessary, and there is no record of transactions. However, it also facilitates criminal financial transactions such as money laundering, financing of terrorism, corruption, and tax evasion. Most of the cash in circulation is in the top two largest denominations, often associated with illicit payments or store of value. 27 36. Recent discussions on the trade-off between financial integrity and privacy for cash could prove useful for CBDC design. Eliminating cash would undermine privacy. Moreover, it is unlikely to improve financial integrity since illicit transactions would presumably migrate to another form of money (McAndrews 2017). Ironically, those seeking anonymity for legitimate purposes might even adopt the parallel money, contributing to its liquidity and attractiveness for criminal use. This pattern may be visible in the adoption of cryptocurrencies. Proposals to do away with highdenomination bills seem more appealing (Rogoff 2016 and references therein). Remaining smalldenomination bills could in part satisfy legitimate preferences for privacy but would pose risks for illicit transactions. 37. Depending on its design, CBDC can strengthen or undermine financial integrity. Financial integrity could be strengthened if authorities impose strict limits on the size of transactions. Alternatively, CBDC can be designed to facilitate effective identity authentication and tracking of payments and transfers. Identities would be authenticated through customer due diligence procedures, and transactions recorded. But unless required by law, users’ information could be protected from disclosure to third parties and governments, while criminals could be deterred by the risk of investigation and prosecution. Although promising on paper, these solutions would have to be further evaluated, and questions answered. For instance, Would users trust the safeguards established to protect their privacy? Would central banks be held responsible for 27 Judson (2017); sample includes Australia, Brazil, Canada, the euro area, Hong Kong SAR, India, Japan, Mexico, Singapore, South Arabia, South Korea, Sweden, Switzerland, Turkey, Russia, the United Kingdom, and the United States. See also Europol (2015). 20 INTERNATIONAL MONETARY FUND CASTING LIGHT ON CENTRAL BANK DIGITAL CURRENCY compliance failures, even if customer due diligence procedures were outsourced? And to what extent could authorities benefit from the ability to scrutinize transaction information for illicit activity in real time? On the other hand, CBDC offering full anonymity and large-value transactions would undermine financial integrity relative to cash and current noncash fund transfer systems. Whatever design is chosen, it should accommodate the implementation of effective AML/CFT measures. Would CBDC Undermine Financial Stability and Banking Intermediation? 38. CBDC could affect financial stability and banking intermediation if it competes with bank deposits. Hence, in what follows we assume CBDC with characteristics similar to those of bank deposits: traceability and protection from loss or theft. In a world of diverse consumers, it is reasonable to assume that some will prefer and adopt CBDC. 39. Two hypothetical scenarios are considered. The first is a tranquil period following the introduction of CBDC. Questions arise such as how banks will respond, what will happen to bank intermediation and funding, and how central banks might react. The second scenario assumes a period of systemic financial stress in a world with CBDC. The key question in this case is what happens to run risk—the potential for a significant shift of deposits from banks to CBDC. Scenario 1: Risk of Disintermediation in Tranquil Times 40. Banks will likely react to the introduction of CBDC, but their ability to defend their business model depends on market power. As some depositors leave banks in favor of CBDC, banks could increase deposit interest rates to make them more attractive.28 But the higher deposit rates would reduce banks’ interest margins. As a result, banks would attempt to increase lending rates, though at the cost of loan demand. 29 The greater banks’ market power, the less demand for credit would contract and the more effectively banks could respond to CBDC by preserving profits (see Box 3). Box 3. Banks’ Response to CBDC—Higher Rates on Deposits and Lending 30 The introduction of CBDC draws deposits away from banks, leading to an upward shift in the deposit supply curve (Figure 5a). Banks counteract some of the impact on their deposit bases by raising deposit interest rates (Figure 5b). Moreover, banks pass part of this deposit rate hike on to their loan interest rates. When banks have more market power in lending (also reflected in the steepness of the demand curve for deposits), they can better insulate their profits by passing the deposit rate hike on to loan 28 Banks could also respond by providing more and better complementary financial services. 29 In addition, central banks could lower policy rates to counter the tighter financial conditions stemming from banks’ higher lending rates, so that the banks’ response to CBDC would be less contractionary for the economy. Moreover, the net impact of CBDC adoption on interest rates will depend on how the central banks introduce the CBDC, where an injection of CBDC via the sale of government bonds could, under specific circumstances, lead to lower rates (Barrdear and Kumhof 2016). 30 This discussion of the reaction of banks to CBDC adoption is based on a model by Agur and others (forthcoming) assuming an oligopolistic market structure in lending markets. INTERNATIONAL MONETARY FUND 21 CASTING LIGHT ON CENTRAL BANK DIGITAL CURRENCY rates. Banks with little market power adjust more aggressively in quantity, exhibiting a larger contraction in deposit and loan volume. Figure 5a. Introduction of CBDC Bank deposit rate Deposit supply with CBDC Account-like CBDC shifts up deposit supply 𝑟2∗ 𝑟1∗ Deposit supply Deposit demand 𝐷1∗ 𝐷2∗ Source: IMF staff Quantity of bank deposits Figure 5b. Effects of CBDC and market power in lending Bank deposit rate ∗ 𝑟2𝑏 Deposit supply with CBDC More market power in lending reduces extent of disintermediation due to CBDC. Instead, deposit rates adjust more. Deposit supply ∗ 𝑟2𝑎 Deposit demand, low market power in lending 𝑟1∗ Deposit demand, high market power in lending Source: IMF staff 22 INTERNATIONAL MONETARY FUND ∗ 𝐷2𝑎 ∗ 𝐷2𝑏 𝐷1∗ Quantity of bank deposits CASTING LIGHT ON CENTRAL BANK DIGITAL CURRENCY 41. Banks could try to replace the deposits that shift to CBDC with other forms of funding. Potential alternatives are commercial paper, bonds, and equity. All are market (wholesale) types of funding. There are three likely implications to such a shift. 42. First, bank funding would become more expensive. Deposit insurance and implicit government guarantees allow banks to fund themselves with deposits at lower cost than with other instruments. 31 Hence, switching away from deposits could result in lower bank profits or higher lending rates to preserve margins. The magnitude of the effect may not be enormous, though, given that most of the value of deposits in most banking systems is uninsured. 43. Second, the introduction of CBDC may affect market discipline in the banking sector. Discipline stems from banks facing higher costs of funding or a drop in deposits as they take more risks (Berger 1991). However, insured depositors do not impose discipline on banks since they do not bear the consequences of the risks they undertake. Following the introduction of CBDC, market discipline could decline (increase) if banks lose more (fewer) uninsured than insured depositors. If market discipline diminishes, banks could take on more risk. 44. Third, bank funding may become less stable. Retail depositors are more stable sources of funding than wholesale depositors (see Huang and Ratnovski 2011; Gertler, Kiyotaki, and Prestipino 2016). If more (fewer) retail depositors prefer CBDC to wholesale depositors, bank funding could become more (less) volatile. In that case, banks might have to hold more liquid assets to meet regulatory requirements or cut back on lending. 45. The extent of disintermediation will be greater among banks in more direct competition with CBDC. CBDC would likely be in greater competition with retail than with wholesale deposits, which require larger-value payments. Thus, banks with a larger share of retail deposits will face tougher competition following the introduction of CBDC and may not be able to raise lending rates to preserve profits. Greater presence of nonbank loan providers would add to competitive pressure on banks. 46. In the event of disintermediation, central banks can respond in several ways, although some are not especially palatable over the medium term and would imply a dramatic step away from central banks’ typical mandates. The central bank could limit the decline in bank deposits and lending by setting limits on individual CBDC holdings or discouraging (such as through fees) convertibility from bank deposits to CBDC. In addition, the central bank could lend the funds diverted from deposits back to banks. This would allow banks to keep lending, although the central bank could require more capital or collateral, with possible implications for the cost and volume of 31 Ueda and di Mauro (2013) estimate that government guarantees have reduced bank funding costs between 60 and 80 basis points in recent decades. Other studies quantify the impact of government guarantees on bank funding costs by analyzing funding cost differentials for banks deemed too big to fail (Tsesmelidakis and Merton 2013; Acharya, Anginer, and Warburton 2016; Kelly, Lustig, and van Nieuwerburgh 2016; Kroszner 2016). INTERNATIONAL MONETARY FUND 23 CASTING LIGHT ON CENTRAL BANK DIGITAL CURRENCY lending. Moreover, the central bank’s balance sheet would grow, it would systematically take on credit risk, and it would have to decide how to allocate funds across banks. 32 47. At the margin, central banks could help maintain the ties between customers and banks by allowing banks to manage CBDC. As discussed earlier, banks could offer wallets to hold and make transactions in CDBC and could offer customer support. Scenario 2: Run Risk in Times of Systemic Financial Stress 48. CBDC may facilitate a generalized run on banks by offering a readily available, safe, and liquid alternative to deposits. 33 But in many cases, this effect will be muted. First, the introduction of CBDC would not facilitate idiosyncratic runs from one bank to another. Such movement can already take place electronically at the click of a button. Second, if a banking crisis coincides with a more general economic (currency or sovereign) crisis, funds will be withdrawn from all local assets, including CBDC. 34 Third, even short of a general crisis, CBDC is unlikely to matter much if very safe and liquid alternatives already exist. These include reserves-only narrow banks or Treasury-only mutual funds 35 or, in some cases, state banks with healthy balance sheets or explicit and credible state guarantees. 36 49. In some cases, CBDC could even help the central bank ease liquidity pressures and thus contain bank runs. 37 The provision of extensive liquidity support is pervasive during banking crises. For example, central banks provided liquidity assistance to banks in almost 96 percent of the 151 crisis episodes studied by Laeven and Valencia (2018). Liquidity can be provided by increasing 32 A situation in which the central bank does not recycle deposits back to banks, but instead takes on the direct role of maturity transformation and intermediation, would mean an even larger departure from the central bank mandate and could give rise to a sovereign–central bank nexus, if the monetary authority is pressured to lend to the government. 33 Bank run models, such as Diamond and Dybvig (1983) and Goldstein and Pauzner (2005) and references therein, emphasize information asymmetries regarding liquidity mismatches on banks’ balance sheets, and liquidity needs of depositors, as the driving forces behind runs. Depositors may also run because of concerns about solvency rather than illiquidity (Calomiris and Gorton 1991). 34 In more general economic crises, depositors may also fear losses in real terms because of high inflation and currency depreciation and may thus attempt to hold foreign assets, leading to capital outflows. In fact, almost 30 percent of banking crises coincide with currency and/or sovereign crises (Laeven and Valencia 2018). Emerging market and developing economies are more susceptible to these twin or triple crises (Caprio and Klingebiel 1996; Laeven and Valencia 2018). Advanced economies are not immune to these episodes as is evident in the recent European sovereign debt crisis (IMF 2011; Brown, Evangelou, and Stix 2017; Sibert 2013). 35 Schmidt, Timmermann, and Wermers (2016) document runs from money market mutual funds following the collapse of Lehman Brothers in 2008, showing that while most prime money market mutual funds experienced outflows, those invested in Treasury bills observed strong inflows “as investors sought the liquidity of the U.S. government market as part of a flight-to-safety.” 36 Barajas and Steiner (2000), focusing on depositor behavior in Colombia, and Mondschean and Opiela (1999), considering Poland, find that state banks are perceived as safer and have an advantage in attracting deposits relative to private banks. However, a large number of studies, focusing on a diverse set of countries, do not offer corroborating evidence (Adler and Cerutti 2015; Hasan, Jackowicz, Kowalewski, and Kozlowski. 2013; Semenova 2007; Das and Ghosh 2004; Ungan and Caner 2004; Hori, Ito, and Murata 2009). 37 The absence of a lender of last resort has been associated with US banking panics in the pre–Federal Reserve era (Calomiris 2008; Gorton and Tallman 2016). 24 INTERNATIONAL MONETARY FUND CASTING LIGHT ON CENTRAL BANK DIGITAL CURRENCY reserves (to replace funding) and cash (to allow banks to meet deposit withdrawals). Having CBDC will not affect the central bank’s ability to increase reserves, since this can already take place electronically. But in geographically vast countries where transporting cash to bank branches and ATMs might be a costly and time-consuming endeavor, CBDC could also facilitate the process of providing liquidity to banks and resolve runs faster. 38 50. Even if the introduction of CBDC increased the risk of systemic bank runs, deposit insurance could alleviate the effects. Liquidity provision needs are smaller in countries with deposit insurance before a banking crisis. 39 Countries adopting CBDC should, therefore, have a deposit insurance scheme to lower the probability of runs. 40 The effectiveness of such schemes in mitigating runs, though, will depend on the credibility of the fiscal backstop and the extent of coverage. Would Monetary Policy Transmission Remain Effective? 51. The introduction of CBDC is unlikely to significantly affect the main channels of monetary policy transmission under plausible CBDC designs. 41 The channels are fourfold: • The basic interest rate channel may be the most affected and could strengthen. Changes in policy interest rates induce households and firms to rebalance investment and consumption between the future and the present, especially if these are exposed to interest-sensitive borrowing and saving instruments. To the extent that CBDC increases financial inclusion—thus access precisely to such instruments—monetary policy transmission could strengthen (Mehrotra and Nadhanael 2016). Gains would be most evident if CBDC were interest bearing. • The bank lending channel could also strengthen. Through this channel, policy interest rates and their expectation affect bank balance sheets and profits—ultimately their creditworthiness and thus their nondeposit funding cost and lending rates. 42 This channel would strengthen if CBDC increased the share of banks’ wholesale funding, as argued earlier. 38 For similar reasons, central banks should not impose aggregate limits on CBDC in circulation. These could induce price distortions as a result of scarcity premiums. These limits could also accelerate runs as there is an attempt to purchase CBDC before others do and the aggregate limit is reached. Note also that the trigger for a run could become more unpredictable; when the run is to CBDC rather than to cash it will be harder for depositors to observe the signal of others’ liquidity needs or information about the bank as there will be no lining up outside branches. 39 In the Laeven and Valencia (2018) sample of crises, median peak liquidity provision was 15.3 percent in countries with deposit insurance; it was 22.4 percent of deposits for countries without it. In addition, there is evidence that uninsured wholesale depositors are more predisposed to runs (Covitz, Liang, and Suarez 2013). 40 In fact, some runs, such as the case of the run on Northern Rock in the United Kingdom in 2007, have been associated with gaps in deposit insurance coverage for small depositors (Goodhart 2011). 41 Similar conclusions are suggested in Meaning and others (2018) and in CPMI (2018). Another unexplored, though interesting, channel is the potential for CBDC to facilitate the resetting of prices and thus weaken transmission. A more remote option requiring attention is for CBDC to allow for interest rates to differ across individuals or regions. 42 See Bernanke (2007) and references therein for a full description of the bank lending channel. Older versions of this channel, originating in a period with higher reserve requirements and credit market segmentation, suggested that a higher supply of reserves increased deposits—loanable funds—and hence bank lending. INTERNATIONAL MONETARY FUND 25 CASTING LIGHT ON CENTRAL BANK DIGITAL CURRENCY • The credit channel is unlikely to be affected much. This channel is related to the one above; policy rates affect asset prices and collateral values of borrowers, thus their creditworthiness and costs of borrowing. 43 However, CBDC should not markedly impinge on these effects. • Likewise, the exchange rate channel—through which changes in policy rates bring about rebalancing between foreign and domestic assets, and a commensurate variation in the exchange rate affecting exports and imports—is unlikely to be affected. 52. This relatively benign view depends on the expectation that central banks would remain in a position to affect market interest rates relevant to the channels above. First, central banks should be able to affect term spreads through communication as before, such as by releasing and discussing their interest rate projections. Second, central banks should be able to retain control of interest rates on reserves. As long as banks demand reserve balances to pay each other— ultimately, as long as banks intermediate payments—the central bank should be able to set their marginal price. 44 This is key, since the price of reserves determines the opportunity cost for banks to lend funds to each other, and by extension the rates set in the much larger money markets. In turn, these affect rates on riskless, and ultimately risky, assets. Clearly, the presence of banks across these markets, as well as arbitrage and lack of market segmentation, is key to transmission. CBDC, however, is not expected to markedly affect any of these conditions under most design scenarios. 53. One scenario, however, would significantly test the standard transmission channels, but could be resolved with a change in operating framework. If banks were no longer involved in intermediating payments, having lost the business to CBDC (or stored value facilities, depending on how these are regulated), demand for reserves would disappear. This scenario resembles the “cashless world” considered by Woodford (2000). In this world, though, monetary policy has the means to remain effective. Woodford argues that “perfect control over overnight rates would still be possible, through adjustments of the rate paid on central bank balances.” In the digital world, the insight translates into paying interest on CBDC. Doing so would put a floor on interest rates if CBDC is provided without limit. Indeed, no one with access to CBDC would lend at a rate below that offered by CBDC, which would remain the safest and most liquid asset available. This is akin to controlling monetary policy by paying interest on reserves when these are in excess of what is demanded by the banking sector for precautionary purposes (referred to as a “floor system”). 43 It could weaken if fintech innovations, beyond just CBDC, reduce information asymmetries inherent in the markups charged to borrowers. 44 Some adjustments may nevertheless be necessary to central banks’ operating frameworks. CBDC is likely to displace cash, but could also partially drain reserves from commercial banks if customers withdraw deposits to hold CBDC. To the extent that banks need the reserves for precautionary purposes, central banks would still be able to replenish these by engaging in liquidity-injecting open market operations. Ultimately, demand for precautionary reserves might actually decrease, because CBDC could attenuate the variance of payment shocks (unlike cash, CBDC does not require lumpy withdrawals from costly visits to ATMs) or increase their predictability. But even if the shape and position of the demand curve for reserves change, central banks should be able to adapt their supply of reserves to stabilize interest rates. In the interim, movements between deposits and CBDC could be volatile and require more frequent liquidity-injecting open market operations—perhaps on a fixed-rate full allotment basis—to stabilize interest rates. A floor system could also be considered to stabilize interest rates, since the demand for liquidity does not need to be accurately forecast. 26 INTERNATIONAL MONETARY FUND CASTING LIGHT ON CENTRAL BANK DIGITAL CURRENCY VI. CENTRAL BANK RESEARCH AND EXPERIMENTS 54. Several central banks, in both advanced and emerging market and developing economies, are considering the pros and cons of issuing CBDC. Table 1 summarizes the jurisdictions in which central banks are (or have been) actively exploring CBDC for retail use based on publicly available information. Table 1. Jurisdictions Where Retail CBDCs Are Being Actively Explored 4546 Australia (on hold) Bahamas Brazil Canada China (and here) Curaçao and Sint Maarten Eastern Caribbean Ecuador (pilot complete) Denmark (rejected) Israel Norway (ongoing) Philippines Sweden United Kingdom (on hold) Uruguay (and here)(pilot) Sources: Central banks or various news sources per hyperlinks above. (Information has not been verified through official channels.) 55. Some sovereigns have issued, or may issue, retail digital currencies, though these are not CBDC because they are not issued by central banks as their liability. For example, the Marshall Islands has discussed launching the SOV, a crypto asset that will become legal tender along with the US dollar, with the apparent intention to raise funds for the government. Likewise, Venezuela is planning to issue the petro, a commodity-backed crypto asset. 56. Central banks are considering CBDC for two main reasons: Declining use of cash in advanced economies and financial inclusion in emerging market and developing economies. These and other objectives mentioned by central banks are summarized in Table 2. None of the central banks surveyed cite seigniorage preservation or monetary policy effectiveness at the zero 45 “Active” means central banks that have convened projects to seriously explore CBDC or that have undertaken pilots. Some central banks have publicly indicated that they are investigating CBDC but have provided little to no detail. These include central banks in Bahrain, Barbados, Egypt, the Euro Area (and rejected), Hong Kong SAR, India, Indonesia, Israel, Jamaica, Korea (and rejected), Lebanon, New Zealand (on hold), Russia, and Switzerland. 46 There is doubt about Senegal’s and Tunisia’s CBDC, which appear to be fiat-collateralized crypto assets. In the case of Senegal’s e-CFA e-currency, the only connection to the central bank seems to be that the e-currency complies with the e-money regulations of the Banque Centrale des Etats de l’Afrique de l’Ouest. In Tunisia, the post office has been operating an e-dinar digital money wallet since 2000, and in 2016 it partnered with Monetas and DigitUs to offer a crypto-powered payment app, but there has been no central bank involvement. INTERNATIONAL MONETARY FUND 27 CASTING LIGHT ON CENTRAL BANK DIGITAL CURRENCY lower bound as rationales for CBDC adoption, with papers by the Bank of Canada emphatically denying the latter. 47 The main rationale for advanced economies seems to be countering the growth of private forms of money (operational risk and monopoly distortions) and reducing costs associated with managing cash (cost efficiency). For example, Sweden points to increasing payment system single-point-of-failure risk associated with diminishing cash usage. In emerging market economies, the main interest in CBDC seems to be fostering financial inclusion by reaching out to the unbanked segments of the population. Only China cites monopoly distortions as a justification. Cost efficiency is also behind China’s investigations, Ecuador’s CBDC issuance (launched in early 2015 though shut down in early 2018), and Uruguay’s six-month trial. 48 Some developing economy central banks have also mentioned reducing the costs and risks associated with the distribution of physical cash. Table 2. Rationales for Exploring CBDCs via Publicly Available Information Diminishing Cash Usage Monopoly Distortions Operational Risks Cost Efficiency Bahamas Financial Inclusion X Canada X China X X X X CBCS X X X ECCB X X X Ecuador Countering derisking X Norway X Senegal Sweden Other X X X Tunisia Uruguay X X X Monetary policy was not cited as a rationale by any of the central banks surveyed. It was not possible to ascertain the rationales, based on publicly available information, for Australia, Bahrain, Denmark, the European Union, Hong Kong SAR, India, Indonesia, Jamaica, South Korea, and Switzerland. Sources: Central banks or various news sources (as indicated in italics) per hyperlinks in Table 1. Information has not been verified through official channels. Note: CBCS = Central Bank of Curaçao and Sint Maarten; ECCB = Eastern Caribbean Central Bank. 47 Engert and Fung (2017) conclude that “reducing the effective lower bound does not provide a compelling motivation to issue CBDC.” 48 Ecuador’s main rationale was to avoid the costs of managing physical dollars in its fully dollarized economy. However, user acceptance was very low, seemingly because of lack of trust in the central bank (White 2018). 28 INTERNATIONAL MONETARY FUND CASTING LIGHT ON CENTRAL BANK DIGITAL CURRENCY 57. Some central banks are reportedly no longer pursuing CBDC. Factors cited include small benefits for central banks and potential disintermediation and bank run risks (Australia, Denmark, European Central Bank, New Zealand, Switzerland). 58. In terms of design, most central banks seem to be contemplating forms of accountbased CBDC, though with various levels of anonymity; some token-based solutions also exist. The Eastern Caribbean Central Bank and the People’s Bank of China could offer both account- and token-based CBDC, with accounts managed by commercial banks and/or other licensed financial institutions. The People’s Bank of China is also considering fully anonymous token-based wallets, albeit with low payment limits. However, other central banks seem to be shying away from such solutions because they would not meet Financial Action Task Force requirements. 49 Most central banks are considering anonymous CBDC when it comes to transacting parties and third parties but not in the case of the government. This is the situation of the People’s Bank of China. (Wallets with greater anonymity, however, would have lower payment limits.) Protecting users against risks of theft and loss was considered important to many of the central banks mentioned. 59. A number of central banks surveyed indicated that CBDC should be available 24 hours a day, 7 days a week, to mimic the accessibility of cash. In this regard, Canada, China, and Sweden are looking into CBDC with offline capability. This would be offered by preloading tokens onto a wallet while online, then validating transactions via encrypted messaging to point-of-sale terminals, similarly to cash cards. In all cases, central banks are considering limits to such transactions. Functionality could be temporarily available during electricity or infrastructure breakdowns but still be susceptible to catastrophic events. 50 Even cash is not particularly resilient in such circumstances, as its distribution and use require electricity (to operate automated teller machines, vending machines, and other cash dispensers), and cash can be destroyed easily in a catastrophe. 60. None of the central banks surveyed are seriously considering interest-bearing CBDC. Central banks appear to be concerned with financial intermediation, lending contraction, and heightened bank balance sheet volatility. Some, such as the Reserve Bank of New Zealand and Uruguay simply assert on principle that CBDC should be fungible. Central banks contemplating token-based CBDC suggest that paying interest would present a technical challenge, though not an insurmountable one. Others raise the hurdle of tracking interest payments for tax purposes. However, Sweden’s proposed CBDC, the e-krona, will have the built-in ability to pay interest if the central bank decides to do so. 61. The central banks surveyed are studying ways of managing and funding CBDC infrastructure. Although marginal costs of managing physical cash are likely higher than those of 49 The Financial Action Task Force is an independent intergovernmental body that develops and promotes policies to protect the global financial system against money laundering, terrorism financing, and the financing of proliferation of weapons of mass destruction. It has set out recommendations for customer due diligence (for instance, identity verification) implementation, record keeping, and suspicious transaction reporting requirements for financial institutions and designated nonfinancial businesses and professions. 50 A recurrence of the 1859 Carrington Event could knock out communications and power for up to a year and render digital money useless. INTERNATIONAL MONETARY FUND 29 CASTING LIGHT ON CENTRAL BANK DIGITAL CURRENCY CBDC, the high up-front fixed costs may favor CBDC adoption in larger economies. Some central banks are exploring the option of building and maintaining CBDC in-house, but most are planning to outsource these onerous tasks, despite the risks involved. Some central banks are contemplating cost-sharing mechanisms. For example, China may make third-party wallet providers bear part of the development costs in return for keeping a cut of any fees charged and benefiting from cross-selling opportunities. The Uruguay pilot appears similar. China is also thinking about charging fees on large transactions, just as for large cash withdrawals. 62. Various central banks surveyed plan to outsource CBDC development, though under close supervision. For example, the Uruguay CBDC pilot turned to the Roberto Giori Company to develop digital notes, IN Switch Solutions for wallets, and Redpagos for “storefront” operations (cashing e-pesos in and out). Tunisia used Monetas for digital notes and DigitUS for wallets, but developed the user interface itself (according to The Tunisian Post). Senegal used eCurrency Mint for the notes, while the Banque Régionale de Marchés provided the wallets and user interface. 51 63. Central banks are for the most part focused on CBDC applications for the domestic economy. Only Canada and China mentioned cross-border issues, but more as complications than opportunities. Canada seemed focused on accessibility by tourists. China cited cross-border capital management. Other projects focus more on intermediating wholesale payments across borders, such as a joint project between the Bank of Canada, the Monetary Authority of Singapore, and the Bank of England. VII. CONCLUSION 64. CBDC could be the next milestone in the evolution of money. The history of money suggests that, while the basic functions of money might not change, the form does evolve in response to user needs. Digitalization of many aspects of economic activity is prompting central banks to seriously consider the introduction of CBDC. 65. CBDC is a digital form of existing fiat money, issued by the central bank and intended as legal tender. It would potentially be available for all types of payments and could be implemented with a variety of technologies. 66. This discussion note introduces a three-step conceptual framework to assess CBDC’s potential to create value both for end users and for central banks. The first step is to identify the criteria with which users evaluate different forms of money. The second step involves establishing the public policy goals of central banks with respect to money. The third step lays out the competitive landscape, comprising existing and evolving forms of money. 67. Overall, the note finds no universal case for CBDC adoption as yet. From the perspective of end user needs, it finds that demand for CBDC will depend on the attractiveness of alternative forms of money. In advanced economies, there may be scope for the adoption of CBDC as a potential replacement for cash for small-value, pseudo-anonymous 51 The International Telecommunications Union is working on standardized terms and conditions for the design of digital fiat currencies, including currencies issued by central banks and managed by private entities. 30 INTERNATIONAL MONETARY FUND CASTING LIGHT ON CENTRAL BANK DIGITAL CURRENCY transactions. But in countries with limited banking sector penetration and inefficient settlement technology, demand for CBDC may well be greater. 68. From a central bank perspective, the case for CBDC is likely to differ from country to country. CBDC may reduce the costs to society that are associated with the use of cash. Moreover, CBDC may improve financial inclusion in cases of unsuccessful private sector solutions and policy efforts. It could also help central banks bolster the security of, and trust in, the payment system and protect consumers where regulation does not adequately contain private monopolies. But regulation and, where possible, novel payment solutions could offer compelling alternatives to a CBDC. 69. For countries that decide to introduce CBDC, appropriate design and policies should help mitigate ensuing risks. Monetary policy transmission is unlikely to be significantly affected and may even benefit from greater financial inclusion. Moreover, though it will not eliminate illicit activity, CBDC may in some situations enhance financial integrity. However, it also entails risks for financial integrity if badly designed. In addition, although CBDC could increase the cost of funding for deposit-taking institutions and intensify run risk in some jurisdictions, design choices and policies can help ease such concerns. Nevertheless, operational and reputational risks arising from malfunctions of the digital infrastructure or cyberattacks are likely to remain as challenges. 70. Looking ahead, the cross-border implications of CBDC raise a multitude of new questions that merit investigation. For instance, from a practical standpoint, how would tourists be able to make payments in a foreign country that has adopted CBDC? Should foreigners have access to CBDC? To what extent would this complicate know-your-customer and AML/CFT compliance, and could standardized information be requested across countries? Would access to CBDC in a reserve currency (such as e-dollars) facilitate currency substitution in countries that have weak institutions? And to what extent might safe-haven flows be encouraged, potentially draining resources from countries that face banking, sovereign, or currency crises? Finally, if CBDC were used for cross-border transactions, how might central banks be required to cooperate? Would they absorb some of the functions of correspondent banks and thus take on additional liquidity, credit, and foreign exchange rate risk—or might tokens be created for cross-border payments among particular central banks, commercial banks, or firms? Research on CBDC should proceed resolutely given that the questions to be explored are deep and difficult and have far-reaching implications. INTERNATIONAL MONETARY FUND 31 CASTING LIGHT ON CENTRAL BANK DIGITAL CURRENCY REFERENCES Acharya, V., D. Anginer, and A. Warburton. 2016. “The End of Market Discipline? Investor Expectations of Implicit Government Guarantees.” https://ssrn.com/abstract=1961656 or http://dx.doi.org/10.2139/ssrn.1961656. Adler, G., and E. Cerutti. 2015 "Are Foreign Banks a Safe Haven? Evidence from Past Banking Crises." IMF Working Paper 15/43, International Monetary Fund, Washington, DC. Agarwal, R., and M. Kimball. 2015. “Breaking through the Zero Lower Bound.” IMF Working Paper 15/224, International Monetary Fund, Washington, DC. Agur, I., A. Ari, and G. Dell’Ariccia. Forthcoming. "Central Bank Digital Currencies: Design Tradeoffs and Implications." IMF Working Paper, International Monetary Fund, Washington, DC. Ali, R. 2018 "Cellular Structure for a Digital Fiat Currency." MIT Digital Currency Initiative. https://dci.mit.edu/digital-fiat-currency. Alvez, M., R. Lluberas, and J. Ponce. 2018. "The Cost of Using Cash and Checks in Uruguay." Unpublished. Arango-Arango, C., Y. Bouhdaoui, D. Bounie, M. Eschelbach, and L. Hernandez. 2018. "Cash Remains Top-of-Wallet! International Evidence from Payment Diaries." Economic Modelling 69 (January): 38– 48. Bennett, B. S. Schuh, and S. Schwartz. 2014. "The 2012 Diary of Consumer Payment Choice." Federal Reserve Bank of Boston Research Data Report. Bagnall, J., D. Bounie, and K. Huynh. 2016. "Consumer Cash Usage and Management: A CrossCountry Comparison with Diary Survey Data." International Journal of Central Banking 12 (4): 1–61. Barajas, A., and R. Steiner. 2000. "Depositor Behavior and Market Discipline in Colombia." IMF Working Paper 00/214, International Monetary Fund, Washington, DC. Barrdear, J. and M. Kumhof. 2016. “The Macroeconomics of Central Bank Issued Digital Currencies.” Bank of England Working Paper 605, Bank of England, London. Bech, M. L., U. Faruqui, F. Ougaard, and C. Picillo. 2018. "Payments Are A-Changin’ but Cash Still Rules." BIS Quarterly Review (March):67–80. Bech, M. L., and R. Garratt. 2017. “Central Bank Cryptocurrencies.” BIS Quarterly Review. September. Berger, A. 1991. “Market Discipline in Banking.” In Proceedings of a Conference on Bank Structure and Competition. Federal Reserve Bank of Chicago. 32 INTERNATIONAL MONETARY FUND CASTING LIGHT ON CENTRAL BANK DIGITAL CURRENCY Bernanke, B. 2007. "Inflation Expectations and Inflation Forecasting." Speech delivered at the Monetary Economics Workshop of the National Bureau of Economic Research Summer Institute, Cambridge, MA. Bordo, M., and A. Levin. 2018. "Central Bank Digital Currency and the Future of Monetary Policy." Monetary Policy and Payments 3:143–78. Brown, M., I. Evangelou, and H. Stix. 2017. “Banking Crises-Bail-ins and Money Holdings.” Central Bank of Cyprus Working Paper 2017-2. Calomiris, C. 2008. “Testimony before the Committee on Oversight and Government Reform.” US House of Representatives, December 9. ———, and G. Gorton. 1991. "The Origins of Banking Panics: Models, Facts, and Bank Regulation." Financial Markets and Financial Crises. University of Chicago Press. Caprio, G. Jr., and D. Klingebiel. 1996. "Bank Insolvencies: Cross-Country Experience." World Bank Policy Research Working Paper Series 1620, World Bank, Washington, DC. Carney, Mark. 2018. "The Future of Money." Speech at the inaugural Scottish Economics Conference, Edinburgh University, March 2. Committee on Payments and Market Infrastructures (CPMI). 2016. “Fast Payments—Enhancing the Speed and Availability of Retail Payments.” CPMI Report 154, Basel. ———. 2018. “Central Bank Digital Currencies.” CPMI Report 174, Basel. Covitz, D., N. Liang, and G. Suarez. 2013. "The Evolution of a Financial Crisis: Collapse of the AssetBacked Commercial Paper Market." Journal of Finance 68 (3): 815–48. Das, A., and S. Ghosh. 2004. "Market Discipline in the Indian Banking Sector: An Empirical Exploration." https://econwpa.ub.uni-muenchen.de/econ-wp/fin/papers/0410/0410020.pdf. Dell'Ariccia, G., V. Haksar, T. Mancini-Griffoli, K. Eckhold, S. Gray, F. Han, G. H. Hong, E. Lundback, H. Oura, H. Poirson, P. Rabanal, and D. Sandri. 2017. "Negative Interest Rate Policies—Initial Experiences and Assessments." IMF Policy Paper, International Monetary Fund, Washington, DC. Diamond, D., and P. Dybvig. 1983. “Bank Runs, Deposit Insurance, and Liquidity.” Journal of Political Economy 91 (3): 401–19. Eichengreen, B. 2018. "Why 'Stable Coins' Are No Answer to Bitcoin's Instability." Project Syndicate, September 11. https://www.project-syndicate.org/commentary/stable-coins-unviablecryptocurrencies-by-barry-eichengreen-2018-09. Engert, W., and B. Fung. 2017. "Central Bank Digital Currency: Motivations and Implications." Bank of Canada Staff Discussion Paper 2017-16, Ottawa. INTERNATIONAL MONETARY FUND 33 CASTING LIGHT ON CENTRAL BANK DIGITAL CURRENCY Esselink, H., and L. Hernández. 2017. "The Use of Cash by Households in the Euro Area." Occasional Paper Series 201, European Central Bank, Frankfurt. Europol. 2015. “Why Cash Is King? A Strategic Report on the Use of Cash by Criminal Groups as a Facilitator for Money Laundering.” https://www.europol.europa.eu/publications-documents/whycash-still-king-strategic-report-use-of-cash-criminal-groups-facilitator-for-money-laundering. Financial Action Task Force (FATF). 2015. “Money Laundering through the Physical Transportation of Cash.” Paris. Flannigan, G., and S. Parsons. 2018. "High-Denomination Banknotes in Circulation: A Cross-Country Analysis." Reserve Bank of Australia Bulletin, March. Fung, B., and H. Halaburda. 2016. “Central Bank Digital Currencies: A Framework for Assessing Why and How.” Bank of Canada Staff Discussion Paper 2016–22, Ottawa. Gertler, M., N. Kiyotaki, and A. Prestipino. 2017. "A Macroeconomic Model with Financial Panics." International Finance Discussion Paper 1219, Federal Reserve Board, Washington, DC. Goldstein, I., and A. Pauzner. 2005. "Demand–Deposit Contracts and the Probability of Bank Runs" Journal of Finance 60 (3): 1293–327. Goodhart, C. 2011. "The Macro-Prudential Authority: Powers, Scope and Accountability." Market Trends 2011-2, Organisation for Economic Co-operation and Development, Paris. Gorton, G., and E. Tallman. 2016. "How Did Pre-Fed Banking Panics End?" NBER Working Paper 22036, National Bureau of Economic Research, Cambridge, MA. Grym, A., P. Heikkinen, K. Kauko, and K. Takala. 2017. “Central Bank Digital Currency.” Bank of Finland Economics Review 5. Gürtler, K., S. Nielsen, K. Rasmussen, and M. Spange. 2017. “Central Bank Digital Currency in Denmark,” Analysis 28, Danmarks National Bank. Habermeier, K., L. Jacome, T. Mancini-Griffoli, C. Baba, T. Bayoumi, J. Chen, G. Dell’Ariccia, S. Gray, M. Ismael, T. Mondino, A. Pescatori, T. S. Sedik, S. Sgherri, H. Tanimoto, K. Ueda, N. Valckx, and F. Valencia. 2013. "Unconventional Monetary Policies—Recent Experiences and Prospects." IMF Policy Paper, International Monetary Fund, Washington, DC. Hasan, I., T. De Renzis, and H. Schmiedel. 2013. “Retail Payments and the Real Economy.” European Central Bank Working Paper 1572, Frankfurt. Hasan, I., K. Jackowicz, O. Kowalewski, and L. Kozlowski. 2013. “Market Disciplines during Crisis: Evidence from Bank Depositors in Transition Countries.” Journal of Banking and Finance 37 (12): 5436–51. 34 INTERNATIONAL MONETARY FUND CASTING LIGHT ON CENTRAL BANK DIGITAL CURRENCY He, D. 2018. “Monetary Policy in the Digital Age.” Finance and Development 55 (2). He, D., K. Habermeier, R. Leckow, V. Haksar, Y. Almeida, M. Kashima, N. Kyriakos-Saad, H. Oura, T. Saadi Sedik, N. Stetsenko, and C. Verdugo-Yepes. 2016. “Virtual Currencies and Beyond: Initial Considerations.” IMF Staff Discussion Note 16/03, International Monetary Fund, Washington, DC. He D., and A. Khan. 2018. “Central Bank Digital Currencies.” In Global Financial Stability Report, Chapter 1, International Monetary Fund, Washington, DC, April. He, D., R. Leckow, V. Haksar, T. Mancini-Griffoli, N. Jenkinson, M. Kashima, T. Khiaonarong, C. Rochon, and H. Tourpe. 2017. "Fintech and Financial Services: Initial Considerations." IMF Staff Discussion Note 17/05, International Monetary Fund, Washington, DC. Hori, M., Y. Ito, and K. Murata. 2009. "Do Depositors Respond Rationally to Bank Risks? Evidence from Japanese Banks during Crises." Pacific Economic Review 14 (5): 581–92. Huang, R., and L. Ratnovski. 2011. "The Dark Side of Bank Wholesale Funding." Journal of Financial Intermediation 20 (2): 248–63. International Monetary Fund (IMF). 2011. “Euro Area Policies: 2011 Article IV Consultation—Staff Report; Public Information Notice on the Executive Board Discussion; and Statement by the Executive Director for Member Countries.” IMF Country Report 11/184, International Monetary Fund, Washington, DC. Jobst, C., and H. Stix. 2017. “Doomed to Disappear: The Surprising Return of Cash across Time and across Countries.” CEPR Discussion Paper 12327, Centre for Economic Policy Research, London. Judson, R. 2017. "The Death of Cash? Not So Fast: Demand for U.S. Currency at Home and Abroad, 1990–2016." Paper presented at the 3rd International Cash Conference, Deutsche Bundesbank, April. Kahn, C., S. Quinn, and W. Roberds. 2014. “Central Banks and Payment Systems: The Evolving Tradeoff between Cost and Risk.” Paper presented at the Norges Bank Conference on the Uses of Central Banks: Lessons from History, June 5–6. Kahn, C. M., F. Rivadeneyra, and T.-N. Wong. 2018. “Should the Central Bank Issue e-Money?” Unpublished. Kahn, C., and W. Roberds. 2009. “Why Pay? An Introduction to Payments Economics.” Journal of Financial Intermediation 18 (1): 1–23. Kelly, B., H. Lustig, and S. van Nieuwerburgh. 2016. “Too-Systemic-to-Fail: What Option Markets Imply about Sector-Wide Government Guarantees.” American Economic Review 106 (6): 1278–319. Kosse, A., H. Chen, M.-H. Felt, V. D. Jiongo, K. Nield, and A. Welte. 2017. “The Costs of Point-of-Sale Payments in Canada.” Bank of Canada Staff Discussion Paper 2017-4, Ottawa. INTERNATIONAL MONETARY FUND 35 CASTING LIGHT ON CENTRAL BANK DIGITAL CURRENCY Kroszner, R. 2016. “A Review of Bank Funding Cost Differentials.” Journal of Financial Services Research 49 (2): 151–74. Krüger, M. 2016. “Pros and Cons of Cash: The State of the Debate.” In 2016, Cash on Trial, edited by C. Beer, E. Gnan, and U. Birchler. SUERF Conference Proceedings 2016/1, SUERF—The European Money and Finance Forum, Vienna, 45–67. Kumhof, M., and C. Noone. 2018. "Central Bank Digital Currencies—Design Principles and Balance Sheet Implications." Bank of England Staff Working Paper 725, London. Laeven, L., and F. Valencia. 2018. "Systemic Banking Crises Revisited." IMF Working Paper 18/206, International Monetary Fund, Washington, DC. McAndrews, J. 2017. “The Case for Cash.” ADBI Working Paper 679, Asian Development Bank Institute, Manila. Meaning, J., B. Dyson, J. Barker, and E. Clayton. 2018. "Broadening Narrow Money: Monetary Policy with a Central Bank Digital Currency." Bank of England Staff Working Paper 724, London. Mehrotra, A., and G. Nadhanael. 2016. "Financial Inclusion and Monetary Policy in Emerging Asia." In Financial Inclusion in Asia, edited by S. Gopalan and T. Kikuchi. London: Palgrave Macmillan, 93–127. Davoodalhosseini, S. 2018. "Central Bank Digital Currency and Monetary Policy." Bank of Canada Staff Working Paper 2018-36, Ottawa. Mondschean, T., and T. Opiela. 1999. "Bank Time Deposit Rates and Market Discipline in Poland: The Impact of State Ownership and Deposit Insurance Reform." Journal of Financial Services Research 15 (3): 179–96. Monetary Authority of Singapore. 2017. “Consultation Paper on Proposed Payment Services Bill.” http://www.mas.gov.sg/News-and-Publications/Consultation-Paper/2017/Consultation-Paper-onProposed-Payment-Services-Bill.aspx. Norges Bank. 2018. "Central Bank Digital Currencies." Norges Bank Paper 1, Oslo. Raskin, M., and D. Yermack. 2016. “Digital Currencies, Decentralized Ledgers, and the Future of Central Banking.” NBER Working Paper 22238, National Bureau of Economic Research, Cambridge, MA. Rogoff, K. 2014. "Costs and Benefits to Phasing Out Paper Currency." NBER Working Paper 20126, National Bureau of Economic Research, Cambridge, MA. ———. 2016. The Curse of Cash. Princeton, NJ: Princeton University Press. Schmidt, L., A. Timmermann, and R. Wermers. 2016. "Runs on Money Market Mutual Funds." American Economic Review 106 (9): 2625–57. 36 INTERNATIONAL MONETARY FUND CASTING LIGHT ON CENTRAL BANK DIGITAL CURRENCY Semenova, M. 2007. "How Depositors Discipline Banks: The Case of Russia." Economics Education and Research Consortium Working Paper Series 07-02e, Kiev. Sibert, A. 2013. “Deposit Insurance after Iceland and Cyprus.” VoxEU.org, April 2. https://voxeu.org/article/deposit-insurance-after-iceland-and-cyprus. Sisak, B. 2011. "What Drives Cash Demand? Transactional and Residual Cash Demand in Selected Countries." Magyar Nemzeti Bank Working Paper 10, Budapest. Sveriges Riksbank. 2017. The Riksbank’s e-krona Project Report 1." September. https://www.riksbank.se/globalassets/media/rapporter/ekrona/2017/rapport_ekrona_uppdaterad_170920_eng.pdf. Tilford, C. 2018. "The Millennial Moment.” In Charts, Financial Times, June 6. https://www.ft.com/content/f81ac17a-68ae-11e8-b6eb-4acfcfb08c11. Tsesmelidakis, Z., and R. Merton. 2013. "The Value of Implicit Guarantees." https://ssrn.com/abstract=2231317 or http://dx.doi.org/10.2139/ssrn.2231317. Ueda, K., and W. di Mauro. 2013. "Quantifying Structural Subsidy Values for Systemically Important Financial Institutions." Journal of Banking and Finance 37 (10): 3830–42. Ungan, A., and S. Caner. 2004. “Depositor Behaviour and Market Discipline in Turkey.” https://ssrn.com/abstract=495122. Wakamori, N., and A. Welte. 2017. "Why Do Shoppers Use Cash? Evidence from Shopping Diary Data." Journal of Money, Credit and Banking 49 (1): 115–69. White. 2018. “The World’s First Central Bank Electronic Money Has Come—And Gone: Ecuador, 2014–2018.” Cato Institute, Washington, DC. Woodford, M. 2000. "Monetary Policy in a World without Money." NBER Working Paper 7853, National Bureau of Economic Research, Cambridge, MA. Yifei, F. 2018. "Some Considerations about Central Bank Digital Currency." YICAI News Service, January 25. https://www.yicai.com/news/5395409.html. INTERNATIONAL MONETARY FUND 37 CASTING LIGHT ON CENTRAL BANK DIGITAL CURRENCY APPENDIX Figure A1.1. Attractiveness to Users of Different Forms of Money Cash Cryptocurrencies Private e-money (by stored value facilities) Traditional bank deposits Improved bank deposits (fast payments & other) CBDC 38 INTERNATIONAL MONETARY FUND CASTING LIGHT ON CENTRAL BANK DIGITAL CURRENCY Table A1.1. Scores by Criteria Cash Traditional bank Cryptocurrencies* deposits Private e-money Bank deposits with (by stored value fast payment facilities) Scalability M: not for M: not for small- L: not for large M: could be H: all transactions large-value or micro-value payments expanded to possible transactions transactions large-value transactions Extra services Interest returns Acceptance Settlement risk L: none H: access to all L: limited services H: integration H: like bank bank services through with social and deposits exchanges financial services L: none L: usually none, L: zero M: interest, yielding though below though could be policy rate offered M: interest M: limits to L: only person to L: only few M: person to H: like stored value change; some business and retailers accept business, business facilities but no merchants business to to business, network limit reject business, point- person to person, of-sale terminals but only within needed network H: none; M: some delay M: lags; struggle M: like bank H: instantaneous in immediate to establish deposits central bank settlement finality for some money (reserves) Theft & loss M: hard to M: can reverse L: hard to recover/ M: like bank M: like bank risk recover/ transaction & claim due to deposits deposits claim; no claim ownership anonymity H: none, a M: deposit H: to the extent M: no deposit M: like bank central bank insurance code is solid, not a insurance, though deposits liability some safeguards L: high energy H: very easy to H: very easy to use, requirements use, cheap cheap H: full anonymity L: not anonymous L: not anonymous cyber risk Default risk liability Transaction L: need to costs physically M: service fees meet Anonymity H: full costs anonymity L: not anonymous Note: Scores are H (high), M (medium), and L (low) to the extent the corresponding criteria are desirable for the end user. For instance, a high mark for transaction costs means that costs are low and thus attractive to users. *Rapidly evolving technological advances could improve scores. INTERNATIONAL MONETARY FUND 39 Committee on Payments and Market Infrastructures Markets Committee Central bank digital currencies March 2018 This publication is available on the BIS website (www.bis.org). © Bank for International Settlements 2018. All rights reserved. Brief excerpts may be reproduced or translated provided the source is stated. ISBN 978-92-9259-142-7 (print) ISBN 978-92-9259-143-4 (online) Foreword The history of central banking began with payment services. Since then payment-related innovation has always been an integral part of central banking. Modern examples include the establishment of systems allowing for immediate interbank gross settlement and the recent increased emphasis on faster retail payment systems. Central bank digital currencies (CBDCs) represent another such potential innovation. This joint report by the Committee on Payments and Market Infrastructures and the Markets Committee provides an initial analysis of CBDCs. It offers a high-level overview of their implications for payments, monetary policy and financial stability. The analysis of the committees reflects initial thinking in this rapidly evolving area and is a starting point for further discussion and research. It also highlights that the issuance of a CBDC requires careful consideration. The Committees thank Klaus Löber (European Central Bank) and Aerdt Houben (Netherlands Bank) and the two Committee working groups for their efforts in preparing this report. Benoît Cœuré Chair, Committee on Payments and Market Infrastructures Central bank digital currencies Jacqueline Loh Chair, Markets Committee iii Contents Foreword ...............................................................................................................................................................................................iii Executive summary ........................................................................................................................................................................... 1 1. Introduction ...................................................................................................................................................................... 3 2. Taxonomy........................................................................................................................................................................... 3 2.1 The money flower .................................................................................................................................................. 4 2.2 Design features ....................................................................................................................................................... 5 3. Payment aspects.............................................................................................................................................................. 7 3.1 General purpose CBDC ........................................................................................................................................ 7 3.2 Wholesale only CBDC ........................................................................................................................................... 8 3.3 Other considerations ............................................................................................................................................ 9 3.4 Key feasibility and operational challenges .................................................................................................. 9 4. Monetary policy aspects ............................................................................................................................................ 10 4.1 Desirability for monetary policy ..................................................................................................................... 10 4.2 Implications for monetary policy implementation and interest rates ............................................ 12 5. Financial intermediation, financial stability and cross-border aspects .................................................... 14 5.1 Role of the central bank .................................................................................................................................... 14 5.2 Banks business models, financial intermediation and markets ......................................................... 15 5.3 Financial stability .................................................................................................................................................. 16 5.4 Cross-border and global dimensions .......................................................................................................... 17 References .......................................................................................................................................................................................... 19 Annex A: Principles of monetary policy implementation ................................................................................................ 21 Annex B: Flow-of-funds representation ................................................................................................................................. 24 Annex C: The impact of CBDC on seigniorage .................................................................................................................... 26 Annex D: Members of the working groups ........................................................................................................................... 27 Committee on Payments and Market Infrastructures ............................................................................................. 27 Markets Committee ............................................................................................................................................................... 28 Central bank digital currencies v Executive summary Interest in central bank digital currencies (CBDCs) has risen in recent years. The Committee on Payments and Market Infrastructures and the Markets Committee recently completed work on CBDCs, analysing their potential implications for payment systems, monetary policy implementation and transmission as well as for the structure and stability of the financial system. Key highlights of the work are:  CBDC is potentially a new form of digital central bank money that can be distinguished from reserves or settlement balances held by commercial banks at central banks. There are various design choices for a CBDC, including: access (widely vs restricted); degree of anonymity (ranging from complete to none); operational availability (ranging from current opening hours to 24 hours a day and seven days a week); and interest bearing characteristics (yes or no).  Many forms of CBDC are possible, with different implications for payment systems, monetary policy transmission as well as the structure and stability of the financial system. Two main CBDC variants are analysed in this report: a wholesale and a general purpose one. The wholesale variant would limit access to a predefined group of users, while the general purpose one would be widely accessible.  CBDC raises old questions about the role of central bank money, the scope of direct access to central bank liabilities and the structure of financial intermediation. Traditionally, central banks have, for various reasons, tended to limit access to (digital) account-based forms of central bank money to banks and, in some instances, to certain other financial or public institutions. By contrast, physical central bank money, ie cash, is widely accessible. This approach has, in general, served the public and the financial system well, setting a high bar for changing the current monetary and financial structure.  Wholesale CBDCs, combined with the use of distributed ledger technology, may enhance settlement efficiency for transactions involving securities and derivatives. Currently proposed implementations for wholesale payments – designed to comply with existing central bank system requirements relating to capacity, efficiency and robustness – look broadly similar to, and not clearly superior to, existing infrastructures. While future proofs of concept may rely on different system designs, more experimentation and experience would be required before central banks can usefully and safely implement new technologies supporting a wholesale CBDC variant.  In part because cash is rapidly disappearing in their jurisdiction, some central banks are analysing a CBDC that could be made widely available to the general public and serve as an alternative safe, robust and convenient payment instrument. In circumstances where the traditional approach to the provision of central bank money – in physical form to the general public and in digital form to banks – was altered by the disappearance of cash, the provision of CBDC could bring substantial benefits. However, analysing whether these goals could also be achieved by other means is advisable, as CBDCs raise important questions and challenges that would need to be addressed. Most importantly, while situations differ, the benefits of a widely accessible CBDC may be limited if fast (even instant) and efficient private retail payment products are already in place or in development.  Although a general purpose CBDC might be an alternative to cash in some situations, a central bank introducing such a CBDC would have to ensure the fulfilment of anti-money laundering and counter terrorism financing (AML/CFT) requirements, as well as satisfy the public policy requirements of other supervisory and tax regimes. Furthermore, in some jurisdictions central banks may lack the legal authority to issue a CBDC, and ensuring the robust design and operation of such a system could prove to be challenging. An anonymous general purpose CBDC would raise further concerns and challenges. Although it is unlikely that such a CBDC would be considered, it would not necessarily be limited to retail payments and it could become widely used globally, including for illegal transactions. That said, compared with the current situation, a non-anonymous CBDC could allow for digital records and traces, which could improve the application of rules aimed at AML/CFT. Central bank digital currencies 1  Issuance of a CBDC would probably not alter the basic mechanics of monetary policy implementation, including central banks’ use of open market operations. CBDC introduces a new type of central bank money whose demand – like cash – would need to be accommodated. CBDC would also not necessarily affect the discretion that central banks have in choosing their monetary policy implementation techniques (eg reliance on purchases of securities or credit operations with banks) as well as the maturity, liquidity and credit risk of their assets. However, if flows into CBDC were to become large and not associated with offsetting declines in physical banknotes, as could be the case in times of financial stress, challenges could arise (such as a need to broaden the assets that the central bank can hold or take on as collateral).  CBDC could enrich the options offered by the central bank’s monetary policy toolkit, eg by allowing for a strengthening of pass-through of policy rate changes to other interest rates or addressing the zero lower bound (or the even lower, effective bound) on interest rates. It is not clear, however, that the current pass-through is anything but adequate. Furthermore, other more conventional tools and policies can to some extent achieve similar outcomes without introducing new risks and challenges (such as implementing negative interest rates on public holdings of a general purpose CBDC). And some of these gains might not arise without discontinuing higher denomination banknotes, which – although helping with AML/CFT requirements – would by itself entail some costs.  Implications are more pronounced for monetary policy transmission and financial markets, especially if a CBDC was to be designed as, or de facto became, an attractive asset. As a liquid and creditworthy asset, a wholesale variant available to institutional investors that would be akin to interest-bearing central bank reserves or reverse repo facilities, yet widely tradeable, could function as a safe asset comparable in nature to short maturity government bills. A general purpose variant could compete with guaranteed bank deposits, with implications for the pricing and composition of banks’ funding.  The introduction of a CBDC would raise fundamental issues that go far beyond payment systems and monetary policy transmission and implementation. A general purpose CBDC could give rise to higher instability of commercial bank deposit funding. Even if designed primarily with payment purposes in mind, in periods of stress a flight towards the central bank may occur on a fast and large scale, challenging commercial banks and the central bank to manage such situations. Introducing a CBDC could result in a wider presence of central banks in financial systems. This, in turn, could mean a greater role for central banks in allocating economic resources, which could entail overall economic losses should such entities be less efficient than the private sector in allocating resources. It could move central banks into uncharted territory and could also lead to greater political interference.  For currencies that are widely used in cross-border transactions, all the considerations outlined above would apply with added force, especially during times of generalised flight to safety. The introduction of a CBDC in one jurisdiction could adversely affect others. Central banks that have introduced or are seeking to introduce a CBDC should consider cross-border issues where relevant.  Any steps towards the possible launch of a CBDC should be subject to careful and thorough consideration. Further research on the possible effects on interest rates, the structure of intermediation, financial stability and financial supervision is warranted. The effects on movements in exchange rates and other asset prices remain largely unknown and also deserve further exploration.  More generally, central banks and other authorities should continue their broad monitoring of digital innovations, keep reviewing how their own operations could be affected and continue to engage with each other closely. This includes monitoring the emergence of private digital tokens that are neither the liability of any individual or institution nor backed by any authority. At this time, the general judgment is that their volatile valuations, and inadequate investor and consumer protection, make them unsafe to rely on as a common means of payment, a stable store of value or a unit of account. 2 Central bank digital currencies 1. Introduction Some central banks have started to consider whether they might, at some stage in the future, issue digital currencies of their own. While providing greater access to digital forms of central bank liabilities is not an entirely new idea (eg Tobin (1985)), the recent debate has been motivated by a number of factors. These include: (i) interest in technological innovations for the financial sector; (ii) the emergence of new entrants into payment services and intermediation; (iii) declining use of cash in a few countries; and (iv) increasing attention to so-called private digital tokens. In response to the growing interest of central banks, the private sector and the public at large, the Committee on Payments and Market Infrastructures (CPMI) and the Markets Committee (MC) conducted complementary studies on the implications of issuing a central bank digital currency (CBDC). This consolidated report is an early contribution to this topic, providing a conceptual analysis of the potential effect of CBDC in three core central banking areas: payments, monetary policy implementation and financial stability. The committees’ work in this area builds on previous work they conducted on the role of central bank money, digital currencies, fast payments, access to central bank services and monetary policy implementation.1 It is complemented by an exploration of possible effects on the structure of the financial system and for financial stability. CBDC raises questions about the role of central bank money, direct access to central bank liabilities and the structure of financial intermediation. Traditionally, central banks have, for various reasons, tended to limit access to (digital) account-based central bank money to banks and, in some instances, to certain other financial or public institutions.2 By contrast, physical central bank money (ie cash) is widely accessible. In some jurisdictions, however, the use of cash is decreasing, with the possibility of its complete disappearance, implying that the public would no longer have wide access to central bank money. Since the traditional approach has, in general, served the public and the financial system well, the bar for changing the current monetary and financial structure is high. The report is organised as follows. Section 2 introduces a taxonomy of CBDC, provides an overview of key design features and describes two variants: a wholesale and a general purpose variant. The two are used as reference cases to analyse the payment system implications in Section 3, as well as the impact on monetary policy implementation and transmission in Section 4. Section 5 discusses the broader implications for the financial system, financial stability risks and cross-border issues. 2. Taxonomy CBDC is not a well-defined term. It is used to refer to a number of concepts. However, it is envisioned by most to be a new form of central bank money. That is, a central bank liability, denominated in an existing unit of account, which serves both as a medium of exchange and a store of value. This would be an innovation for general purpose users but not for wholesale entities. Central banks already provide digital money in the form of reserves or settlement account balances held by commercial banks and certain other financial institutions at the central bank. This mix of new and already existing forms of central bank money makes it challenging to precisely define what a CBDC is. In fact, for purposes of analysing what may change, 1 See also CPSS (2003), CPMI (2012), CPMI (2014), CPMI (2015) and CPMI (2016a, 2016b). 2 In the early days of central banking, it was fairly common to offer accounts not just to banks but also to non-banks (see eg Reichsbank (1926) and Bank of England (1963)). However, starting in the 20th century, central banks have tended to progressively restrict access by non-banks. In recent years, access has been granted to some critical financial market infrastructures (FMIs), such as central counterparties (CCPs), mainly for financial stability purposes. Moreover, some central banks have provided access to liquidity-absorbing instruments, such as central bank bills and reverse repos, to a broader set of counterparties than banks. Central bank digital currencies 3 it is easier to define a CBDC by highlighting what it is not: a CBDC is a digital form of central bank money that is different from balances in traditional reserve or settlement accounts.3 2.1 The money flower To get greater clarity, it is useful to put CBDC in the context of other types of money. Graph 1 presents a taxonomy of money in the form of a Venn-diagram referred to as the money flower (Bech and Garratt (2017)). The version here focuses on the combinations of four key properties: issuer (central bank or other); form (digital or physical); accessibility (widely or restricted); and technology (token- or account-based).4 Money is typically based on one of two basic technologies: tokens of stored value or accounts (Green (2008) and Mersch (2017a)). Cash and many digital currencies are token-based, whereas balances in reserve accounts and most forms of commercial bank money are account-based. A key distinction between token- and account-based money is the form of verification needed when it is exchanged (Kahn and Roberds (2009)). Token-based money (or payment systems) rely critically on the ability of the payee to verify the validity of the payment object. With cash the worry is counterfeiting, while in the digital world the worry is whether the token or “coin” is genuine or not (electronic counterfeiting) and whether it has already been spent.5 By contrast, systems based on account money depend fundamentally on the ability to verify the identity of the account holder. A key concern is identity theft, which allows perpetrators to transfer or withdraw money from accounts without permission.6 Identification is needed to correctly link payers and payees and to ascertain their respective account histories. Digital central bank money is at the centre of the money flower. The taxonomy distinguishes between three forms of CBDCs (the dark grey shaded area). Two forms are token-based and the other is accountbased. The two token-based versions differ first and foremost by who has access, which, in turn, depends on the potential use of the CBDC. One is a widely available payment instrument that is primarily targeted at retail transactions but also available for much broader use.7 The other is a restricted-access digital settlement token for wholesale payment and settlement transactions. Below they are referred to as (central bank) general purpose token and (central bank) wholesale token. The account-based version envisages the central bank providing general purpose accounts to all agents in the jurisdiction. While the scale would be of a different magnitude, the technology to do so is arguably currently available. The novelty would be the decision to implement such accounts. 3 Reserves and settlement accounts are available in most jurisdictions to “monetary policy counterparties”, ie financial institutions that are directly relevant for monetary policy implementation, such as deposit-taking entities, which are generally already granted access to central bank deposit and lending facilities. In some jurisdictions, account holders may comprise a broader group and include non-monetary counterparties (eg treasury, foreign central banks or certain financial markets infrastructures (FMIs)). Some central banks are considering widening access. CBDC would further expand access to digital central bank money but not to central bank lending facilities. 4 Accessibility distinguishes between money that is available everywhere to everyone and money that is restricted to certain agents or jurisdictions. 5 Double-spending is a potential problem for digital tokens. There is a risk that a payer could try to use the “same” token on two different transactions. 6 The incident that occurred in February 2016 at the central bank of Bangladesh is an example of false verification based on compromised credentials. CPMI (2017b) presents a strategy to counter fraud in wholesale payment systems. In general, safeguarding against unauthorised access or tampering of account histories is of utmost importance. If someone maliciously breaks into the trusted intermediary hosting all the account balances, they can in principle tamper or modify any account balances at will. CPMI and IOSCO (2016) provides guidance on cyber-resilience for financial market infrastructures. 7 It is common to divide payments into retail and wholesale segments. Retail payments are relatively low-value transactions, in the form of, for example, cheques, credit transfers, direct debits and card payments. By contrast, wholesale payments are largevalue and high-priority transactions, such as interbank transfers. The distinction might become less relevant in a world with CBDCs. In fact, depending on its design, a widely available CBDC could also be used for wholesale transactions. 4 Central bank digital currencies The money flower: a taxonomy of money Graph 1 Notes: The Venn-diagram illustrates the four key properties of money: issuer (central bank or not); form (digital or physical); accessibility (widely or restricted) and technology (account-based or token-based). CB = central bank, CBDC = central bank digital currency (excluding digital central bank money already available to monetary counterparties and some non-monetary counterparties). Private digital tokens (general purpose) include crypto-assets and currencies, such as bitcoin and ethereum. Bank deposits are not widely accessible in all jurisdictions. For examples of how other forms of money may fit in the diagram, please refer to the source. Source: Based on Bech and Garratt (2017). 2.2 Design features In addition to the four core properties highlighted above, there are other design features that will determine how a CBDC may serve as a means of payment and a store of value. These choices will have implications for payments, monetary policy and financial stability. The most important CBDC design options identified to date are listed below. Table 1 provides a comparison of properties across existing and potential new forms of central bank money. Availability. Currently, access to digital central bank money is limited to central bank operating hours, traditionally less than 24 hours a day and usually five days a week.8 CBDCs could be available 24 hours a day and seven days a week or only during certain specified times (such as the operating hours of largevalue payment systems). CBDC could be available permanently or for a limited duration (eg it could be created, issued and redeemed on an intraday basis). 8 The introduction of faster or instant payment systems in an increasing number of jurisdictions has led a number of central banks to reconsider the time during which access to digital central bank money is available, with some moving toward availability 24 hours a day seven days a week for central bank money settlement of fast retail payments (see CPMI (2016b) and Bech et al (2017)). Central bank digital currencies 5 Key design features of central bank money Table 1 Existing central bank money Central bank digital currencies Cash Reserves and settlement balances 24/7 availability    () () Anonymity vis-à-vis central bank   ()  () Peer-to-peer transfer   ()  () Interest-bearing  () () () () Limits or caps   () () () General purpose token accounts Wholesale only token  = existing or likely feature, () = possible feature,  = not typical or possible feature. Anonymity. Token-based CBDC can, in principle, be designed to provide different degrees of anonymity in a way that is similar to private digital tokens. 9 A key decision for society is the degree of anonymity vis-à-vis the central bank, balancing, among other things, concerns relating to money laundering, financing of terrorism and privacy. Transfer mechanism.10 The transfer of cash is conducted on a peer-to-peer basis, while central bank deposits are transferred through the central bank, which acts as an intermediary. CBDC may be transferred either on a peer-to-peer basis or through an intermediary, which could be the central bank, a commercial bank or a third-party agent. Interest-bearing. As with other forms of digital central bank liabilities, it is technically feasible to pay interest (positive or negative) on both token- and account-based CBDCs. The interest rate on CBDC can be set equal to an existing policy rate or be set at a different level to either encourage or discourage demand for CBDC.11 Both non-interest bearing and interest bearing accounts could be used for retail or wholesale payment transactions. The payment of (positive) interest would likely enhance the attractiveness of an instrument that also serves as a store of value. Limits or caps. Different forms of quantitative limits or caps on the use or holdings of CBDC are often mentioned as a way of controlling potentially undesirable implications or to steer usage in a certain direction. For example, limits or caps could make a CBDC less useful for wholesale rather than retail payments. At present, such limits or caps on holdings/use are most easily envisioned in non-anonymous account-based systems.12 9 For example, bitcoin allows transactions to be (pseudo) anonymous. While all bitcoin transactions are publicly recorded using the payer’s and the payee’s public addresses, very much like e-mail addresses, these addresses do not necessarily reveal the true identity of users. A person sending bitcoin to a public address thus need not reveal his/her true identity to the recipient (counterparty anonymity) or to other users (one form of third-party anonymity). Recent innovations may allow even more anonymity than in the original bitcoin design. 10 Bech and Garratt (2017) focus on the transfer mechanism (centralised or decentralised) rather than on the token- or accountbased technology. Money is either exchanged in a decentralised manner known as peer-to-peer (ie transactions occur directly between the payer and the payee without the need for a central intermediary) or in a centralised manner relying on the services of one or more third parties. Tokens are often transferred peer-to-peer. 11 Moreover, rates could be differentiated. For example, if accounts were linked to individual persons or entities, the CBDC rate could vary by counterparty, amount held in the account or some other characteristic, in a way that is similar to the current central bank practice of extensive use of differentiated interest rates on deposits held by non-monetary counterparties. 12 The proper functioning of the payment system, however, implies one-to-one convertibility of CBDC with respect to reserves and banknotes (Fung and Halaburda (2016)). Not facilitating one-to-one convertibility would lead to an exchange rate between different types of central bank money, breaking the unity of the currency. However, some have proposed allowing this unity to break under certain circumstances. For example, Agarwal and Kimball (2015) propose abandoning one-to-one convertibility as a way of allowing a floating exchange rate between cash and commercial bank deposits and thus eliminating the effective lower bound. Abandoning convertibility between CBDC and reserves would similarly lead to a floating exchange rate between CBDC and commercial bank deposits. 6 Central bank digital currencies The different combinations of features mean that there are many potential CBDC variants. The two variants analysed below – one with restricted access for wholesale payments and one with wide access for general purposes (either token- or account-based) – are presented for conceptual clarity purposes only; they are by no means exhaustive. 3. Payment aspects The introduction of a general purpose or a wholesale only CBDC could bring a number of potential benefits to payment, clearing and settlement systems, but it could also pose several risks and challenges. In deciding the case for CBDCs, central banks should compare them with existing or enhanced payment, clearing and settlement solutions. And they would need to consider the impacts on other parts of their remit – most importantly monetary policy and financial stability (analysed in the next two sections). 3.1 General purpose CBDC One rationale for introducing CBDC in a jurisdiction could be to provide a safe, central bank instrument, especially should the use of cash decline significantly. Over the past decades, technological developments have significantly improved the convenience and efficiency of digital forms of private sector payment instruments compared with central bank paper money (ie banknotes). In Sweden, these developments have led to an absolute decline in the amount of cash in circulation. The Riksbank is investigating whether an e-krona would provide the general public with continued access to central bank money and increase the resilience of the payment system (Skingsley (2016) and Sveriges Riksbank (2017)).13 While specifics will vary according to a country’s circumstances and economic conditions, these payment-related motivations for issuing CBDC appear at this time not to be compelling for most jurisdictions. The growing use of electronic means of payment has generally not yet resulted in a substantial reduction in the demand for cash (Graph 2).14 The rationale for considering a central bank replacement for, or supplement to, cash thus may appear less compelling (CPMI (2017a)). The efficiency gains for retail payment purposes may also be less material. In many countries, current retail payment solutions are convenient, efficient and reliable, and have earned public trust and confidence over time. Going forward, technology will likely offer even more opportunities to enhance convenience, increase safety, lower overall costs and further improve resilience. A number of jurisdictions have already adopted or are in the process of addressing public demand for faster and more convenient approaches to payments that are also compatible with new digital and mobile technologies. Some are already providing real-time or near real-time settlement and close to 24/7 availability. One exception is perhaps cross-border retail payments, which are generally slower, less transparent and more expensive than domestic retail payments (CPMI (2018)). Some argue that CBDC could also reinforce the resilience of a country’s retail payment systems. They argue that should payments in private sector infrastructures be disrupted due, say, to technical problems or because a bank providing credit transfers was under stress, households and businesses could still make digital payments via CBDC, something especially important if cash had (largely) disappeared. On a related note, CBDC could reduce the concentration of liquidity and credit risk in payment systems (Dyson and Hodgson (2016)). However, one could, of course, also achieve operational resilience through the diversity afforded by multiple payment systems, although this could be difficult to achieve given the network effects and economies of scale present in payment systems. In addition, continued availability and use of physical currency could help ensure even greater resilience by providing an instrument that is more immune to disruptions to electric power and telecommunication networks resulting from natural or man-made 13 Cash use has declined to the point where a growing number of merchants no longer accept cash and most bank branches have eliminated cash processing (Skingsley (2016)). 14 Unfortunately, internationally comparable data are not available on the actual use of cash, only for cash in circulation. Central bank digital currencies 7 disasters. Having said that, in those jurisdictions where the general public is abandoning cash, this is not an alternative. Card payments and cash demand, change 2007–161 As a percentage of GDP 1 Graph 2 The start of an arrow represents 2007 data while the end represent 2016 Source: Bech et al (2018). In this context, one could also consider the implications of not issuing CBDC. One is the potential for private digital tokens to more widely displace central bank money in transactions. Retail customers could face more credit and liquidity risks, relative to central bank liabilities, from exposure to either private issuers of digital tokens or from a lack of issuer. At this time, their volatile valuations and inadequate investor and consumer protection make private digital tokens unsafe to rely on as a common means of payment and a stable store of value or unit of account. Overall, while carefully monitoring the development and potential uses of new technologies, central banks are likely to continue to emphasise the need for improving the efficiency and speed of private systems. 3.2 Wholesale-only CBDC In terms of wholesale markets, the main argument made is that settlement systems for financial transactions could be made more efficient – in terms of operational costs and use of collateral and liquidity – and more secure by using wholesale CBDC. Introducing a wholesale CBDC that is comparable to traditional central bank reserves into interbank payment systems could potentially improve efficiency and risk management in settlement (see CPMI (2017a)). If complemented by direct participation of non-banks in the settlement process, gains could further increase, including through facilitating the use of new technologies for asset transfers, authentication, record-keeping, data management and risk management. Payments and (cash legs of) securities transactions settled in CBDC, instead of through facilities hosted by commercial banks or other service providers, could help reduce counterparty credit and liquidity risks in the financial system. It could also help central banks monitor financial activity. To meet evolving needs from financial markets and to ensure an overall stable and sound financial system, a number of central banks have been conducting experiments involving CBDC and its related underlying technology (in particular DLT). Early experimentation, however, has not shown significant benefits for wholesale payments. The design of an infrastructure using such new technology would look similar to the one currently in place in terms of legal, operational and security requirements. Doubts remain regarding the maturity of the technology and the size of efficiency gains associated with the use of DLT. Moreover, changes could imply expanded – direct or indirect – access to a central bank account with new counterparties, which could be difficult to control. That said, technologies and related possible designs are evolving quickly and central banks will need to continually assess whether introducing CBDC (potentially incrementally) in this area could be useful. 8 Central bank digital currencies 3.3 Other considerations In addition to more efficient and safer payments and settlement systems, CBDC could come with additional benefits. Given that a CBDC can allow for digital records and traces, it could improve the application of rules aimed at anti-money laundering and countering the financing of terrorism (AML/CFT), and possibly help reduce informal economic activities. These gains may, however, be small in that the formal payment system, especially if there were to be a traceable CBDC, would not necessarily be the main conduit for illicit transactions and informal economic activities. There are also costs. Commercial banks could lose a valuable interface with their consumers given that in some CBDC designs the “know-your-customer” function could fall to the central bank. Central banks would have to take on a much larger role in this field, with associated costs. Central banks could also be called upon to provide information to tax and other authorities (eg for judicial matters). Moreover, they would have to manage privacy and anonymity issues stemming from the insights obtained from private transactions. More generally, central banks might have to deal with many requests and customers, including some now excluded, for which they are not necessarily well equipped (although some of these challenges may be mitigated or avoided by careful design). Another argument is that a CBDC could improve financial inclusion. In some countries, a sizable portion of the population does not participate in the formal financial system and could thus miss out on associated benefits. A CBDC, however, does not necessarily alleviate all the constraints to access; for some segments of the population, barriers to the use of any digital currency may be large, and the preference for trusted alternatives, such as cash, is strong. In addition, a CBDC may allow for better real-time data on economic activity but such gains are already largely achievable with existing payments data. A more persuasive argument is that a CBDC may help to maintain a direct link between central banks and citizens (especially where cash use is diminishing), which could help foster the public’s understanding of central banks’ roles and need for independence (Mersch (2017b)). 3.4 Key feasibility and operational challenges Even if CBDCs were deemed desirable, initial exploration and experimentation have identified a number of legal, technical and operational issues that central banks and other relevant parties must consider before an instrument can be deemed suitable for wide-scale use. In some countries, there are legal considerations. Not all central banks have the authority to issue digital currencies and expand account access, and issuance may require legislative changes, which might not be feasible, at least in the short term. Other questions include whether a CBDC is “legal tender” (ie a legally recognised payment instrument to fulfil financial obligations) and whether existing laws pertaining to transfers of value and finality are applicable.15 Central banks would also have to take account of AML/CFT concerns and requirements if they were to issue CBDC. Issuing a CBDC that does not adequately comply with these and other supervisory and tax regimes would not be advisable. To date, it is not clear how AML/CFT requirements can be implemented practically for anonymous forms of CBDC. Forms of CBDC that can be easily transferred across borders or used offshore are especially likely to present significant challenges in this respect. As such, the reputational risk to the central bank from a general purpose CBDC must be considered. The use of central bank and commercial bank deposits typically provides some level of privacy (for individual banks and agents, respectively), while the use of cash provides anonymity to all users. The appropriate degree of privacy, as also judged by society, is a challenge in a digital environment. For CBDC, 15 Existing laws are typically written broadly for direct physical transfer or for a central entity (“banks”) to accept instructions and modify a ledger. In a CBDC based on DLT, multiple entities could modify a set of distributed ledgers. Other legal issues, such as the timing of the discharge of obligations and liability for errors and unauthorised payments, may also be relevant. Central bank digital currencies 9 the appropriate degree of privacy of the currency would need to be considered carefully, which could entail difficult public policy design choices for a central bank. Cyber-security is currently one of the most important operational challenges for central bank systems and the financial industry more generally. Cyber-threats, such as malware, and fraud are risks for nearly every payment, clearing and settlement system. They pose, however, a particular challenge for a general purpose CBDC, which is open to many participants and points of attack. Moreover, the potential effect of fraud could be more significant because of the ease with which large amounts could be transferred electronically. Robust mitigation methods of cyber-risk would therefore be a prerequisite for CBDC issuance. More generally, the robustness of possible new technologies in ensuring a sound risk management framework is uncertain. Because central bank services are essential to the smooth functioning of an economy, very robust requirements for reliability, scalability, throughput and resilience are necessities. Central banks therefore typically have very rigorous operational requirements for their systems and services. Some of the proposed technologies for issuing and managing CBDC (such as DLT) are still relatively untested, and even the private sector is in the early phase of developing and applying DLT for commercial use.16 Many questions surrounding operational risk management and governance need to be answered before deployment can be envisioned. This may especially be the case for countries at earlier stages of financial infrastructure development. 4. Monetary policy aspects The consequences of CBDC issuance for the implementation and transmission of monetary policy are directly related to how wide access to CBDC is and whether it is attractively remunerated. Monetary policy arguments for issuing CBDC include potential strengthening of the pass-through of the policy rate to money markets and deposit rates, and helping to alleviate the zero (or effective) lower bound constraint. These arguments should be considered carefully. It is not clear that the pass-through of the policy rate needs strengthening and introducing a CBDC may also bring new risks to monetary policy. In addition, existing tools can, in many cases, achieve the same objectives. Since digital central bank money is already available to monetary counterparties and some non-monetary counterparties, as discussed in other sections, this section refers only to the monetary policy aspects introduced by wider access to CBDC.17 4.1 Desirability for monetary policy Wider digital access to the central bank may strengthen the pass-through of the policy rate to money and lending markets. Monetary policy implications are likely more pronounced if CBDC emerges as an attractive asset to hold. The crucial design features that determine the extent to which CBDC may function as such include the rules regulating its access by different types of agent, its availability beyond intraday use and whether it is interest-bearing, and at what rate (Box A). Only if it combines these choices, would it be a new and liquid central bank liability likely to have an impact on the channels of transmission of policy rates to the money market and beyond. 16 Any CBDC need not necessarily be implemented using some form of DLT; theoretically more traditional centralised technologies may suffice. The pros and cons of using DLT in general, eg as regard to scalability, confidentiality and resilience, is an area of ongoing research that is outside the scope of this report. 17 Besides the fact that digital central bank money is already provided to monetary counterparties, and merely changing the technology behind the provision of funds is thus of limited significance, there are three reasons for this delineation. First, while central banks may need to adjust the quantity of money provided to monetary counterparties to control short-term interest rates, the demand for central bank money held by non-monetary counterparties (eg treasury, foreign central banks or certain FMIs) is more typically just accommodated. Second, there may be good reasons for central banks to provide digital central bank money on different terms (remuneration, settlement hours, individual quantitative limits and anonymity) to (various) monetary and non-monetary counterparties. Third, while monetary counterparties have some access to intraday and overnight credit (ie reserve balances may turn negative), non-monetary counterparties typically do not. Similarly, CBDC balances may not turn negative. 10 Central bank digital currencies Box A Features of CBDC, demand and the degree of substitution with other financial assets The way in which access to CBDC is granted implies that substitution effects will affect different types of financial asset. CBDC accessible to individuals and designed as a non-interest bearing, retail payment instrument might primarily substitute for cash (eg token-based CBDC) and commercial bank deposits (eg account-based CBDC). CBDC that pays interest and is readily transferable would likely be attractive to professional financial market participants (eg cash pools and asset managers). It may substitute for money market instruments, such as government bills, reverse repos, central bank bills and FX swaps, and be a liquid and credit risk-free asset facilitating final settlement. CBDC accessible to nonresidents may substitute for internationally-used banknotes, bank deposits and international reserve assets. Substitution may theoretically be limited by imposing individual quantitative limits in normal times, eg access could be conditional upon a commercial bank account to which payments are redirected in case this upper limit is surpassed, to try to curb demand. Substitution effects will be importantly influenced by whether a CBDC is non-remunerated (as is cash), whether it pays interest at an unchangeable or adjustable rate and whether that rate might possibly move with the policy rate, and, if so, at a spread that is constant or varying. Moreover, rates could be differentiated. A substantially lower interest rate on CBDC holdings exceeding, say, the amount covered by deposit insurance schemes would reduce their attractiveness in normal times. These and other design features will influence the demand for CBDC. If designed with limited attractiveness, the substitution effects in normal times may be moderate, and so will be the effects on monetary policy transmission (as well as any structural effects on the financial system). Of course, in times of stress, central banks are unlikely to want to directly control the quantity of CBDC because they would want to maintain one-to-one convertibility with respect to reserves and banknotes. Even if purposefully designed to be primarily a payment vehicle, CBDC may still end up functioning like a store of value in unforeseen ways under certain circumstances. In times of financial stress, domestic (retail) investors are likely to consider CBDC attractive relative to bank deposits, with many possible side effects, including for financial stability (see section 5). And, if granted access, residents in high-inflation countries may turn to CBDC issued by a lowinflation country (as they do nowadays with cash).  An application of overall quantitative limits to CBDC may potentially disrupt payment systems, giving rise to an exchange rate between different types of central bank money. Such issues may not occur in the case of individual quantitative limits. However, the aggregate of individual limits could in theory produce a binding overall limit in certain situations.  The one-to-one convertibility between CBDC, banknotes and reserves means that the central bank can only control their joint quantity. While the central bank can, in principle, steer the overall quantity of central bank money outstanding through liquidity-injecting and liquidity-absorbing open market operations, the holders of central bank money jointly determine its composition, as they are free to convert one type of liability into another. Commercial banks face a similar issue in not being able to directly control the quantity of their retail deposits. This illustrates that means of payment cannot be directly quantitatively controlled but are rather indirectly influenced by their design features and adjustments in other items. Central banks already face this issue in the provision of banknotes, reserves and deposits for a relatively small number of non-monetary counterparties (see Annex A). Traditionally, central banks passively and elastically accommodate the demand for banknotes and deposits held by non-monetary counterparties to steer the quantity of reserves. This is a necessary condition for implementing monetary policy and it would apply with equal force to CBDC. In particular, a CBDC attractively remunerated compared with other interest rates could affect holdings by institutional investors of other liquid, low-risk instruments (such as short-term government bills and repos backed by sovereign collateral).18 If institutional investors could hold such an instrument without limits, the interest rate on it would help establish a hard floor under money market rates, which is arguably useful.19 An interest-bearing general purpose variant could also make pass-through more direct. If households considered a CBDC to be an alternative to commercial bank deposits, banks would have less scope for independently setting the interest rate on retail deposits. For example, banks would find it harder not to 18 Note also that this refers to the general collateral (cash-driven) segment of repo markets, not to the “specials” (collateraldriven) segment. 19 Duffie and Krishnamurthy (2016), who do not explicitly mention CBDC as a possible instrument, argue that the dispersion of rates that is related to imperfect pass-through signals a social cost. Central bank digital currencies 11 increase deposit rates in tandem when the central bank was raising the CBDC rate. As such, a change in the policy rate could be more directly transmitted to bank depositors (possibly with an intermediation margin, given costs and credit risks). To the extent that an attractively remunerated CBDC reduced currency substitution, which is a possibility in some countries, pass-through more generally could be enhanced, including with respect to domestic prices. In principle, negative rates on central bank liabilities could provide the monetary stimulus needed in extreme circumstances. Proponents have suggested that issuance of CBDC could serve to alleviate the zero lower bound if it came along with a reduced desire for cash holdings (eg Goodfriend (2016) and Dyson and Hodgson (2016)). Relatedly, some argue that having a substitute for cash in the form of (interest-bearing) CBDC makes the discontinuation of higher denomination banknotes easier to achieve (Rogoff (2016) and Bordo and Levin (2017)). 20 There are, however, important caveats and counter-arguments. The degree to which key market rates move in conjunction with the policy rate appears satisfactory for most central banks. Whether the passthrough to money markets, for example, is impeded in material ways is not clear (Potter (2017)). Moreover, it is not clear whether one should expect bank deposit rates to respond immediately to policy rate changes. The spreads between the policy rate and retail rates represent compensations for various risks and transaction costs, including for services that are implicitly cross-subsidised (commercial banks provide a broader range of services to retail investors than any CBDC would). More generally, retail depositors tend to be less price-sensitive than wholesale investors. And, the stickiness of retail deposits allows commercial banks to perform more easily their maturity, credit risk and liquidity transformation roles in the economy. In practice, the lack of a one-to-one response to policy rate hikes and cuts does not represent a challenge as long as central banks have appropriate control over financial conditions. Banks take into consideration a wider range of factors than simply the policy rate in the pricing of their retail deposits, including longer-term rates that encompass credit and liquidity risk premia (ie they look at the relevant investment opportunities). That said, the presence of an attractive CBDC would put pressure on commercial banks to raise their retail deposit rates to avoid losing retail funding. At the same time, some doubt that additional tools would strengthen the central bank’s ability to achieve its objectives (eg Bindseil (2016)). Moreover, even if pass-through warrants strengthening, there are other conventional tools, such as central bank bills, time deposits and standing reverse repo facilities (Box B) that can accomplish the same objective. Box B Central bank bills, time deposits and standing repo facilities as alternatives to CBDC Strengthening the pass-through of the policy rate to money market rates also could be achieved by the central bank supplying liquidity-absorbing instruments to non-bank money market participants. The latter includes reverse repo facilities, time deposits and central bank bills. Central banks have significant expertise and experience in the use of such tools. Considering the pros and cons of these alternatives, there are two key differences between offering liquidity-absorbing instruments and CBDC to money market participants:  CBDC can be used as intraday liquidity by its holders, whereas liquidity-absorbing instruments cannot achieve the same, or can do so only imperfectly. At the moment, there is no other short-term money market instrument featuring the liquidity and creditworthiness of CBDC. The central bank would thus use its comparative advantage as a liquidity provider when issuing CBDC.  Although the quantity of CBDC can be influenced by its design features, it cannot be fully controlled. By contrast, liquidity-absorbing instruments can be auctioned off in fixed quantities. While a CBDC could carry a negative rate, this may not address effectively the zero lower bound if higher denomination banknotes were not simultaneously abolished (eg Pfister (2017)). More generally, considering political economy consequences, it is uncertain how deeply negative rates may work in 20 12 Also, some have argued that CBDC could enhance the effectiveness of quantitative easing, given that monetary counterparties would no longer have to intermediate when the central bank conducted asset purchases (eg bonds would be swapped for riskfree CBDC) instead of dealing in credit-risky commercial bank deposits, possibly strengthening any portfolio rebalancing effects. Central bank digital currencies practice, (McAndrews (2017)). Finally, weaker demand for cash does not imply the need for a CBDC. In fact, monetary policy can still remain effective even without cash (Woodford (2000)). On balance, it is not clear that there is a strong basis at this time to issue a CBDC for the purpose of enhancing the efficacy of monetary policy transmission. 4.2 Implications for monetary policy implementation and interest rates The presence of CBDC would have a limited impact on monetary policy implementation – that is, how central banks use their balance sheets to control short-term interest rates (for a review see Annex A). While a central bank would need to accommodate demand for CBDC, flows into CBDC would drain the amount of reserves in the system in exactly the same way as flows into banknotes and central bank deposits held by non-monetary counterparties (eg the treasury, foreign central banks or financial market infrastructures (FMIs)) currently do. In a corridor system, all flows in and out of CBDC need to be compensated through liquidity-injecting and liquidity-absorbing open market operations (OMOs) to keep the desired amount of reserves.21 In a floor system, only when CBDC inflows drained reserves to the point where they became scarce would the central bank need to undertake additional liquidity-injecting OMOs. Therefore CBDC does not alter the basic “mechanics” of monetary policy implementation (see further Annex B for a flow-of-funds representation). Demand for CBDC would just be another factor to consider for policy responses to be consistent with continued control over short-term interest rates. There are two practical implications, though. First, depending on the degree of substitution, a larger balance sheet may be needed to implement monetary policy, as agents substitute physical cash, commercial bank deposits and other safe assets for CBDC. Second, the overall volatility of autonomous factors could be affected, which, in turn, may affect their predictability.22 While likely requiring larger balance sheets, central banks would still have discretion in choosing the assets they hold to accommodate the demand for CBDC, just as they have for banknotes. Theoretically, assets can be made up of outright holdings of any kind or collateralised lending to monetary counterparties on any terms and conditions. 23 Subject to the overall supply of various types of asset and changes thereof, the additional duration, liquidity and credit risk stemming from accommodating the demand for CBDC is thus determined by the central bank itself, as is the case with banknotes. Demand for CBDC may be volatile on a daily basis, as inflows and outflows result from payments between CBDC and non-CBDC holders. Whether this leads to higher overall volatility depends on the correlations with other factors.24 If volatility proves particularly high, central banks can be forced to operate through a floor system. Whether the quality of liquidity forecasting is hampered depends on the predictability of daily flows in and out of CBDC. The overall effects of CBDC on the (term) structure of interest rates are very hard to predict and will depend on many factors. To attract demand, short-term government paper and overnight repos with treasury collateral might have to provide some yield pickup with respect to a wholesale-oriented 21 Under a corridor system, the (marginal) CBDC remuneration rate should not exceed the policy rate. Otherwise, monetary counterparties would have an incentive to trade their excess overnight funds with CBDC holders instead of trading them among themselves. Monetary counterparties with temporary liquidity deficits would need to bid up overnight rates, causing shortterm interest rates to exceed the policy rate. Under a floor system, the marginal CBDC rate should not exceed the rate of remuneration of reserves placed at the central bank’s deposit facility. 22 CBDCs are considered an autonomous factor for monetary policy implementation for two reasons. First, from the viewpoint of the day-to-day steering of the central bank’s balance sheet to control short-term interest rates, daily fluctuations in the demand for CBDC are an exogenous factor, even though CBDC would be an endogenous factor within the broader monetary policy framework. Second, even if CBDC was introduced, the amount of digital central bank money held by monetary counterparties (reserves) would still be crucial for control over short-term interest rates. 23 As central bank credit to monetary counterparties is collateralised, a widening of collateral eligibility may be necessary to accommodate banks’ increased recourse to credit facilities to compensate for the loss of funding due to CBDC inflows (Annex B). 24 In the case of a corridor system, this may necessitate more frequent liquidity-injecting and liquidity-absorbing OMOs, higher reserve requirements with averaging provisions or wider tolerance bands around reserve targets to steer liquidity conditions. Central bank digital currencies 13 remunerated CBDC. This means that the short end of the sovereign yield curve may end up above the CBDC rate. Contrary to the hard floor that the wholesale CBDC variant may put under money market rates, the general purpose variant is likely to put only a soft floor under retail deposit rates given the lower price sensitivity of retail depositors and switching costs. At the same time, depending on the specific assets held to accommodate the issued CBDC, central banks would probably need to engage in various kinds of maturity, liquidity and credit risk transformation. How these two forces balance out in terms of various interest rates across assets classes and maturities is difficult to predict. More generally, the implications of a CBDC relative to other instruments are likely to depend on each jurisdiction’s specific operating environment. Also, since operating environments may change in the future, monetary policy cost-benefit analyses related to CBDC may need to be revisited periodically. 5. Financial intermediation, financial stability and cross-border aspects Whether or not to introduce a CBDC depends on an assessment of many fundamental issues that go beyond the impact on the payment system and monetary policy transmission and implementation. In this section, topics warranting further investigation are explored. 5.1 Role of the central bank A fundamental matter raised by CBDC issuance relates to the appropriate roles – in financial intermediation and the economy at large – of private financial market participants, governments and central banks. With CBDCs, there could be a larger role for central banks in financial intermediation. As the demand for CBDC grows, and if holdings of cash do not decline in lockstep, central banks might need to acquire (or accept as collateral) additional sovereign claims and, depending on size, private assets (eg securitised mortgages, exchange-traded funds and others). If demand becomes very large, central banks may need to hold less liquid and riskier securities, thereby influencing the prices of such securities and potentially affecting market functioning. Central banks may also need to provide substantial maturity, liquidity and credit risk transformation at times to both banks and markets. Since central banks could assume more important roles, they could have a larger impact on lending and financial conditions. Given that all this could challenge the two-tier banking system, structural implications need to be understood better before CBDC issuance can take place. A greater role for central banks in credit allocation entails overall economic losses if central banks are less efficient than the private sector at resource allocation (eg as it impedes the efficient use of decentralised knowledge in society (Hayek (1945)). It is doubtful, for example, that, from the perspective of an efficient allocation of credit, a centralised approach involving outright holdings of corporate securities would be preferred to a decentralised approach based on banks and other private actors granting loans to corporations and investing in securities. From an infrastructure perspective, central banks would have to decide on the design of the appropriate technology, create the required infrastructure and governance and manage this new form of money. This could lead to large operational demands and associated (upfront) costs, with the possible creation of new risks. There could also be changes to market liquidity and interlinkages. If the demand for CBDC exceeded the decline in the demand for cash and/or reserves, larger outright holdings of CBDC could hamper market functioning if they reduced the free-floating share of outstanding bonds. While a CBDC would by itself be very liquid, it could result in reduced liquidity and increased “specialness” in collateral (repo) markets. The depth of repo and short-term government bill markets could decline as demand was redirected to wholesale market use of CBDC. While the central bank could step in on the demand side of these markets, it would need to broaden its holdings to match its increasing liabilities. This expanded role of central banks in wholesale markets could also reduce interbank activity and the price discovery role of these markets. 14 Central bank digital currencies Coordination issues between the central bank and the government debt management office might occur and central banks’ operations could become more challenging (Greenwood et al (2014)). By having to passively accommodate the demand for CBDC, the central bank could potentially introduce volatile demand for government debt. Related questions include which part of the public sector is best suited to issue a country’s short-term public debt and determine the maturity profile of the consolidated public debt. If CBDC replaced a large portion of bank deposits, central bank demand for government securities could be large, which might then affect sovereign debt markets. More broadly, a larger balance sheet could present challenges as it reduced the role of the market in price setting. Such a reduction could lead to allocative distortions and tie up higher-quality assets. This could, in turn, adversely affect the functioning of collateral markets. All of this would have implications for financial stability. Depending on design, central banks’ seigniorage income could also be affected (see Annex C). Relatedly, if CBDC was interest bearing, the central bank would be directly exposed to stakeholders that might at times exert pressures to raise interest rates. Applying differentiated rates (eg by amount held or counterparty) could also be necessary for effective monetary policy implementation but this might prove to be technically difficult (eg on token-based CBDC). It could also lead to arbitrage as well as being controversial (eg a CBDC rate for households below the rate of remuneration on excess reserve balances). 5.2 Banks business models, financial intermediation and markets The issuance of CBDC would have implications for the structure of payment markets. To the extent that a CBDC would further open up payments to non-banks, commercial banks would stand to see their payment-related income streams eroded by increased competition. Private sector FMIs, such as securities settlement systems and possibly central counterparties for securities trades, might be affected by the issuance of wholesale CBDC.25 While such developments may be far off – because of the many legal, technical and market coordination challenges involved – market participants and authorities would need to be alert, as indirect or unintended consequences might occur. A general purpose CBDC could have a large impact on financial intermediation patterns. The consequence of a larger central bank balance sheet could be a withdrawal of funding to commercial banks. For example, a flow of retail deposits into a CBDC could lead to a loss of low-cost and stable funding for banks, with the size of such a loss in normal times depending on the convenience and costs of the CBDC. Banks could try to prevent a loss of deposits by raising interest rates or seek funding to replace such outflows, eg through wholesale funds and term deposits, which would likely be more costly.26 This could lead some banks to raise spreads and increase transaction fees in order to maintain profitability. Depending on existing market structures, including the importance of retail versus wholesale funding, banks might have to shrink their balance sheets, with possible adverse consequences.27 Commercial banks’ business models would also have to adapt. Services that are currently crosssubsidised by deposits would need to become viable on a stand-alone basis. The contours of institutions undertaking the liquidity, credit risk and maturity transformation no longer performed by banks are not clear. If liquidity in financial markets were to decline and credit and term spreads were to rise, there could 25 New applications of technology could allow participants to interact directly with a synchronised securities ledger to add, verify and report transactions, with activity to be accelerated, at least theoretically, to real-time settlement. In such a vision, central counterparties might no longer be necessary to guarantee trades between execution and settlement. A wholesale CBDC might be considered by some central banks to be part of their toolkit to improve settlements. Nonetheless, many legal, technical and market coordination challenges would need to be addressed first. Multilateral coordination and governance over such arrangements would also likely be necessary. And regulatory authorities would insist on prudent management. 26 Furthermore, alternative means of funding are subject to uncertainties. First, the issuance of bonds by banks is contingent upon placement with investors, which may face some obstacles during times of market stress. Second, any increase in refinancing via the central bank is usually limited by the amount of assets that can be pledged as collateral with the central bank. Third, regulatory constraints may further limit the options available to compensate for the loss of deposits. 27 Annex B contains a flow-of-funds analysis illustrating stylised static balance sheet adjustments of key sectors of the economy upon the introduction of an interest-bearing and widely accessible CBDC. Central bank digital currencies 15 be adverse repercussions for the economy.28 More generally, the implications of a shrinkage of commercial bank balance sheets and activity are very hard to assess and require further analysis. A CBDC attracting significant demand as an asset to hold, may also change the structure and functioning of funding markets, affecting banks and corporations. Issuers of money market instruments and borrowers in repo markets would see more competition because a CBDC would substitute for such claims. Those who issue claims bought by the central bank to accommodate demand for CBDC would gain. Overall, there might also be a collateral upgrade for private balance sheets if central banks end up holding some less liquid and lower-rated assets to accommodate the issuance of CBDC. 5.3 Financial stability Issuance of CBDC raises questions that are similar to those relating to narrow banking or full-reserve money, as analysed by several academics and critics of current monetary systems. Proponents claim that narrow banking could make the overall financial system safer because it limits the scope for commercial banks’ operations. Although narrow banking raises many questions in its own right, the introduction of a CBDC does not necessarily entail the same restrictions.29 While difficult to anticipate, the possibility that banks could try to offset the higher cost of funding by engaging in riskier forms of lending to restore profitability could create financial stability risks. While such risks would have to be compared with those associated with other (unconventional) monetary policy tools, and combined with the potential adverse economic impact of reduced lending (Stevens (2017)), there could be more, rather than less, financial stability risk. In terms of wholesale markets, some (eg Greenwood et al (2016)) argue that the provision of a safe and ultra-liquid asset may help reduce rollover risks and excessive maturity transformation, potentially improving financial stability. However, whether a CBDC leads to these benefits relative to other tools is uncertain (Box C). Arguably, the most significant and plausible financial stability risk of a general purpose CBDC is that it can facilitate a flight away from private financial institutions and markets towards the central bank. Faced with systemic financial stress, households and other agents in both advanced and emerging market economies tend to suddenly shift their deposits towards financial institutions perceived to be safer and/or into government securities. Of course, agents could always flee towards the central bank by holding more cash. But a CBDC could allow for “digital runs” towards the central bank with unprecedented speed and scale. Even in the presence of deposit insurance, the stability of retail funding could weaken because a risk-free CBDC provides a very safe alternative. Depending on the context, the shift in deposits could be large in times of stress. A crucial element in such system-wide shifts is the stronger sensitivity of depositors to the actions of others. The more other depositors run from weaker banks, the greater the incentive to run oneself. If CBDC were available, the incentives to run could be sharper and more pervasive than today, as the CBDC would be the favoured destination, especially if deposits were not insured in the first place or deposit insurance was (made more) limited.30 Whereas weaker banks could experience a run, even stronger banks could face withdrawals in the presence of CBDC. 28 There are also questions in terms of microprudential regulation and supervision. Would, for example, regulatory requirements, such as capital and liquidity adequacy, and supervision of banks, need to be adapted? 29 Narrow banking and CBDC differ in two ways. First, under CBDC residents hold direct claims on the central bank, whereas under narrow banking residents hold commercial bank money that is fully backed by central bank reserves or sovereign claims. Second, CBDC could coexist with commercial bank money, whereas narrow banking proposals envision no private money creation. Benes and Kumhof (2012) and Cochrane (2014), which represent examples of recent calls for narrow banking, also review historical precedents, such as the Chicago Plan of the 1930s. Bacchetta (2017) critically reviews such a proposal in the case of Switzerland. 30 Although with a lower stock of demand deposits commercial banks might be less prone to retail runs, runs in recent times have been initiated by other (wholesale) creditors, which would become more important. 16 Central bank digital currencies It would be difficult to stem runs under such conditions, even when providing large lender of last resort facilities. Changes in the interest rate that applies to CBDC are unlikely to succeed when agents seek safety at almost any price. Imposing quantitative limits, difficult at any time owing to various forms of evasion, could create price deviations between types of central bank money (“discounts”), negating the principle of money being exchangeable at par and hampering the conduct of monetary policy. Box C CBDC, rollover risk and financial stability A secular rise of institutional cash pools and a stronger desire among investors for secured forms of financing have increased the demand for highly liquid and safe instruments, which cannot be met by bank deposits (Pozsar (2011)). This has led to a “near-money premium” in wholesale markets, ie yields on short-term, liquid instruments that are significantly lower compared with those of slightly longer tenors or higher credit risk. This, in turn, can incentivise agents to fund longer-term assets with short-term liabilities (eg repo or commercial paper), with associated rollover risks that could adversely affect financial stability. Central banks may have a role in reducing these risks by providing non-banks with an attractive money-type instrument. As argued by some (eg Stein (2012)), the augmented supply of safe assets may force market participants to scale back their funding of longer-term assets with short-term wholesale borrowing. If less liquid and riskier money market instruments (eg commercial paper) lost some of their near-money premium, the incentives faced by issuers for maturity, liquidity and credit risk transformation could be weakened. Whether a CBDC would materially reduce rollover risks, however, is uncertain. Moreover, increased issuance of short-term debt by the government can also reduce the near-money premium, with possibly associated benefits. Moreover, central banks have other conventional tools at their disposal that could serve a similar purpose (Box B).  The term “institutional cash pool” refers to large, centrally managed, short-term cash balances of global non-financial corporations and institutional investors, such as asset managers, securities lenders and pension funds.  See eg Greenwood et al (2016) or Carlson et al (2016) for further analysis. Another way the near-money premium expresses itself is when short-term government bills and short-term repos with sovereign collateral trade significantly below the overnight index swap (OIS) rate and the policy rate.  For example, long positions in government bonds financed mostly in repo markets (leveraged fixed income strategies employed by hedge funds) could be unwound as collateral chains between institutional investors and money market funds are disintermediated (Pozsar (2011) and Singh (2016)). 5.4 Cross-border and global dimensions For currencies widely used in cross-border transactions, many of the considerations outlined above would apply with added force. In normal times, there would be many complications should non-residents be allowed to hold and transact in CBDC. Distinctions between residents and non-residents and domestic and foreign transactions could become largely symbolic. For example, it could be more difficult to apply AML/CFT requirements because of a lack of formal powers over intermediaries involved in token-based CBDC distribution. Similarly, if foreign banks and FMIs (and even other central banks) were able to purchase, receive or otherwise hold “domestic” CBDC, legal and operational issues could arise. For example, a foreign entity could use the domestic CBDC to back or otherwise provide the functional equivalent of “offshore” accounts and payment services denominated in the domestic currency. Further, the more anonymous the instrument and the more decentralised the transfer mechanism was, the greater the opportunity for cross-border activity, arbitrage and concealed transactions would be, with related reputational risks for the central bank. A CBDC available cross-border could, in some economies, increase substitution away from the domestic currency, which could make monetary aggregates unstable and alter the choice of monetary instruments. Even during normal times, CBDC could come with first-mover advantages and economies of scale and other externalities. In terms of market share, if CBDCs were introduced by jurisdictions with international currencies, they could reinforce existing costs and benefits, including externalities. Similarly, CBDC could change the nature of global liquidity and safe asset provision. Also, and especially if introduced in a sudden and unexpected manner, CBDC could, in some situations, lead to large capital movements and related exchange rate and other asset price effects. In addition, countries might face challenges in preparing for Central bank digital currencies 17 what would happen if other central banks were to introduce CBDC. More generally, disturbances could easily occur. The cross-border and global dimensions of CBDCs available to non-residents could be especially pronounced during times of generalised flight to safety. Under such conditions, exchanging a CBDC for an international currency could potentially enable faster deleveraging in capital markets. If CBDCs accelerated flights from risk, deleveraging pressures could manifest themselves in the form of tight funding conditions and sharp movements in foreign exchange markets. 18 Central bank digital currencies References Agarwal, R and M Kimball (2015): “Breaking through the zero lower bound”, IMF Working Papers, no WP/15/224. Bacchetta, P (2017): “The sovereign money initiative in Switzerland: an economic assessment”, CEPR Discussion Papers, no 12349. Bank of England (1963): “Origin of the branches.” /media/boe/files/quarterly-bulletin/1963/branches-of-the-boe.pdf. https://www.bankofengland.co.uk/- Bech, M and R Garratt (2017): “Central bank cryptocurrencies”, BIS Quarterly Review, September, pp 55–70. Bech, M, Y Shimizu and P Wong (2017): “The quest for speed in payments”, BIS Quarterly Review, March, pp 57–67. Bech, M, U Faruqui, F Ougaard, and C Picillo (2018): “Payments are a-changin’ – but cash still rules”, BIS Quarterly Review, March, pp 67-80. Benes, J and no WP/12/2012. M Kumhof (2012): “The Chicago Plan revisited”, IMF Working Papers, Bindseil, U (2014): “Monetary policy operations and the financial system”, Oxford University Press. ——— (2016): ”Evaluating monetary policy operational frameworks”, paper presented at the Economic Policy Symposium at Jackson Hole. Bordo, M and A Levin (2017): “Central bank digital currency and the future of monetary policy”, NBER Working Papers, no 23711. Carlson, M, B Duygan-Bump, F Natalucci, B Nelson, M Ochoa, J Stein and S Van den Heuvel (2016): “The demand for short-term, safe assets and financial stability: some evidence and implications for central bank policies”, International Journal of Central Banking, vol 12, no 4, December, pp 307–333. Cochrane, J (2014): “Toward a run-free financial system”, University of Chicago, working paper. Committee on Payment and Settlement Systems (CPSS) (2003): The role of central bank money in payment systems, August. Committee on Payments and Market Infrastructures (CPMI) (2014): Non-banks in retail payments, September. –––– (2012): Innovations in retail payments, May. –––– (2015): Digital currencies, November. –––– (2016a): A glossary of terms used in payments and settlement systems, October. –––– (2016b): Fast payments – Enhancing the speed and availability of retail payments, November. –––– (2017a): Distributed ledger technology in payment, clearing and settlement: An analytical framework, February. –––– (2017b): Discussion note - Reducing the risk of wholesale payments fraud related to endpoint security - consultative document, September. –––– (2017c): Statistics on payment, clearing and settlement systems in the CPMI countries - Figures for 2016 (preliminary version), October. –––– (2018): Cross-border retail payments, March. CPMI and International Organization of Securities Commissions (IOSCO) (2016): Guidance on cyber resilience for financial market infrastructures, June. Central bank digital currencies 19 Duffie, D and A Krishnamurthy (2016): “Adapting to changes in the financial market landscape”, paper presented at the Economic Policy Symposium at Jackson Hole. Dyson, B and G Hodgson (2016): “Digital cash: why central banks should start issuing electronic money”, Positive Money. Fung, B. and H. Halaburda (2016): “Central Bank Digital Currencies: A Framework for Assessing Why and How.” Bank of Canada Staff Discussion Paper No. 2016-22. Goodfriend, M (2016): “The case for unencumbering interest rate policy at the zero bound”, paper presented at the Economic Policy Symposium at Jackson Hole. Green, E (2008): “Some challenges for research in payment systems” in The Future of Payment Systems, eds A Haldane, S Millards and V Saporta, Routhledge, Milton Park. Greenwood, R, S Hanson, J Rudolph and L Summers (2014): “Government debt management at the zero lower bound”, Hutchins Centre on Fiscal & Monetary Policy at Brookings, Working Papers, no 5. Greenwood, R, S Hanson and J Stein (2016): “The Federal Reserve’s balance sheet as a financial-stability tool”, paper presented at the Economic Policy Symposium at Jackson Hole. Hayek, F (1945): “The vol 35, no 4, pp 519–30. use of knowledge in society”, American Economic Review, Kahn C and W Roberds (2009): “Why pay? An introduction to payments economics”, Journal of Financial Intermediation 18(3), January, pp 1–23. McAndrews, J (2017): “The case for cash”, ADBI working paper, no 679. Mersch, Y (2017a): “Digital base money: an assessment from the ECB’s perspective”, speech at the Bank of Finland, 16 January. ——— (2017b): “Why Europe still needs cash”, Contribution for Project Syndicate. Pfister, C (2017): “Monetary policy and digital currencies: much ado about nothing?” Bank of France, Working Papers, no 642. Potter, S (2017): “Money markets at crossroads: policy implementation in times of structural change”, remarks at the University of California, Los Angeles. Pozsar, Z (2011): “Institutional cash pools and the Triffin dilemma of the US banking system”, IMF Working Papers, no WP/11/90. Reichsbank (1926): Die Reichsbank, 1901–1925, Druckerei der Reichsbank. Rogoff, K (2016): The curse of cash, Princeton University Press. Singh, M (2016): Collateral and financial plumbing, Risk Book, 2nd impression. Skingsley, C (2016): “Should the Riksbank issue e-krona?”, speech at FinTech Stockholm, 16 November. Stein, J (2012): “Monetary policy as financial-stability regulation”, Quarterly Journal of Economics, 127, February, pp 57–95. Stevens, A (2017): “Digital currencies: threats and opportunities for monetary policy”, National Bank of Belgium, Economic Review, June. Sveriges Riksbank (2017): The Riksbank’s e-krona project – Report 1, September. Tobin, J (1985): “Financial innovation and deregulation in perspective”, Cowles Foundation Papers, no 635. Woodford, M (2000): “Monetary policy in a world without money,” International Finance, vol 3, no 2, pp 229–60. 20 Central bank digital currencies Annex A: Principles of monetary policy implementation This Annex provides a short overview of the general principles of monetary policy implementation, namely the use of the central bank’s balance sheet to achieve its operational target. This target, which can be controlled by the central bank on a day-to-day basis, is highly relevant to the fulfilment of its mandate (Bindseil (2014)). Typically, central banks use an overnight rate as their operational target. The financial institutions that are directly relevant to this operational target and its transmission to money markets are the central bank’s monetary counterparties. To achieve their operational target, central banks need to ensure that the value of attracting or trading away overnight funds from monetary counterparties equals the operational target. Two operational regimes are typically used for this purpose: a corridor and a floor system. In a corridor system, central banks apply two interest rates to reserves: (i) up to a limited amount (depending on reserve requirements), the policy rate is applied; and beyond that (ii) a substantially lower deposit rate is paid.31 Monetary counterparties may access an overnight lending facility at a higher rate. Central banks continuously need to ensure via open market operations (OMOs) that the overall amount of reserves equals the overall limit amount at which the policy rate applies. Central banks can increase flexibility in fulfilling this requirement by applying: (i) a band at which the policy rate applies instead of a limit; or (ii) the minimum required amount of reserves averaged over a maintenance period. Central banks must forecast the demand for liquidity in order to be prepared to inject (or drain) the right quantity of reserves. This involves projecting day-to-day changes in autonomous factors – that is, all the balance sheet items outside of the direct control of the central bank’s monetary policy implementation function that affect the amount of reserves. The difference between the policy and the deposit rate provides an incentive for monetary counterparties to trade overnight funds among themselves, on a secured or unsecured basis. Abstracting from possible balance sheet and collateral costs, such transactions take place close to the policy rate. Thus, the policy rate becomes the marginal value of attracting or trading away overnight funds from monetary counterparties, while the overall amount of reserves can be relatively small. This enables central banks to run a relatively lean balance sheet. This means a balance sheet that is only slightly larger than banknotes outstanding, limiting the intermediary role of the central bank (Graph A1). Under a floor system, central banks ensure that the marginal value of attracting or holding overnight funds from monetary counterparties equals the deposit rate. With substantial excess reserves, the marginal use for monetary counterparties of holding additional reserves is to earn the deposit rate (Graph A2). The deposit rate thereby becomes the de facto policy rate. To achieve this, monetary outright holdings must exceed the original liquidity deficit, ie the liquidity needs caused by net autonomous factors. Liquidity forecasting is less important because day-to-day fluctuations in the amount of reserves do not change the marginal value of attracting or holding overnight funds (with monetary counterparties). In both operational regimes, flows into non-monetary deposits, that is digital central bank money held by non-monetary counterparties (eg the treasury, foreign central banks or FMIs) and banknotes result in a drain of reserves. In a corridor system, such flows need to be compensated by liquidity-injecting OMOs. In a floor system, such flows only need to be compensated if the liquidity surplus becomes insufficient and rates begin to rise above the deposit rate (monetary outright holdings threaten to fall below the original liquidity deficit). In practice, flows into banknotes are limited by the carrying cost of cash, making banknotes relatively inconvenient as a store of value. Flows into non-monetary deposits are typically limited by price 31 Under zero reserve regimes, such as that of the Bank of Canada, the central bank charges a higher policy rate on negative balances (ie loans) and pays a lower deposit rate on positive balances. Under this system, required reserves are not necessary and the overall limit amount at which the policy rate applies can be zero. Central bank digital currencies 21 disincentives beyond certain specified amounts, also making non-monetary deposits relatively unattractive as a store of value. Such price disincentives are often applied to limit the central bank’s intermediary role. Different central banks put varying weights on this principle, however, and apply different price disincentives and access conditions to non-monetary deposits. A stylised balance sheet of the central bank after the introduction of CBDC is depicted in Graph A3, reflecting the demand for CBDC and its increased assets holdings. Corridor system without CBDC Graph A1 Floor system without CBDC Graph A2 22 Central bank digital currencies Central bank balance sheet with CBDC Central bank digital currencies Graph A3 23 Annex B: Flow-of-funds representation A stylised flow-of-funds analysis illustrates qualitatively how sectoral balance sheets and the implementation and transmission of monetary policy may be affected by the introduction of a general purpose CBDC. The more CBDC is perceived by economic agents to be an attractive asset, the larger will be the substitution effects discussed below. The balance sheets considered are those of: i. Households (retail). It is assumed that households hold real assets (RA), retail deposits at commercial banks (DEP) and banknotes (BAN). Furthermore, they invest in corporate/government and bank bonds (B + BB) and money market fund shares (FS) if their liquid funds exceed deposit guarantee schemes’ coverage. They finance themselves through retail mortgage loans (RML) provided by commercial banks and their own funds or equity (E). ii. Corporations/government. It is assumed that corporations and the government fund themselves via bank loans (L) and bonds (B) as well as money market instruments (MM). This sector holds real assets (eg public infrastructure, corporate facilities) and liquidity buffers in the form of cash pool participations (CPP). iii. Banks (monetary counterparties). Funding takes place by accepting retail deposits, by issuing money market instruments (eg secured funding via repos or unsecured funding via commercial paper) and bank bonds and by drawing on central bank credit facilities. These instruments fund purchases of government and corporate bonds, loans to corporates, retail mortgages to households and holdings of central bank reserves (RES). iv. The central bank. The liability side of the central bank’s balance sheet consists of banknotes held by households and reserve balances held by banks. On the asset side, the central bank has outright holdings of corporate, government and (covered) bank bonds and provides credit to banks, therewith implementing monetary policy. The introduction of CBDC opens up a number of channels that affect patterns of financial intermediation in the economy (see the bold, red font balance sheet items in Table B1).32 First, households may substitute banknotes for CBDC (CBDCa), which prompts a change on the central bank’s liability side. Second, households may substitute retail deposits for CBDC (CBDCb) by making payments from retail deposits to CBDC accounts. To effect such payments, banks request the central bank to debit reserves held by them and credit the CBDC accounts. In order to ensure that reserves stay at the required level to implement monetary policy, the central bank buys bonds or provides additional credit to banks. 33 The main question is how large these flows are likely to be and how financial market participants that attract or lose funding will adjust their behaviour. What assets will the central bank hold against the CBDC inflows? Will the financial market participants that lose funding raise funds elsewhere or will they deleverage? Table B1 shows qualitatively one of the many possible outcomes. The central bank accommodates CBDC inflows by increasing its lending to monetary counterparties and outright holdings of bonds. The banks use the central bank’s funds to compensate for the lost retail deposits (CBDCb). In this highly restrictive scenario, there is only a shift in intermediation and no impact on the real assets held by corporates/governments and households (ie no deleveraging and/or leveraging). Instead, the central bank intermediates between households, on the one hand, and banks and corporates/government, on the other. 32 Further substitution effects could be induced as money market funds switch holdings of money market instruments (eg reverse, repos, commercial paper or treasury bills) for CBDC. These effects are omitted from the analysis for ease of exposition. 33 Hence, it is assumed that the central bank either implements monetary policy through a corridor or a floor system with a minimum amount of excess liquidity, consistent with keeping short-term rates close to the deposit rate. 24 Central bank digital currencies CBDC and the structure of the financial system: a flow-of-funds analysis1 Table B1 Households (retail) Real assets Retail deposits CBDC Banknotes Bonds (for investment) (Money market) fund shares RA1 Equity DEP – CBDCb Retail mortgage loans E RML CBDCa + CBDCb BAN – CBDCa B1 + BB1 FS Corporations/government Real assets RA2 Loans Cash pool participation CPP Corporate/government bonds MM instruments L B1 + B2 + B3 MM1 Banks (monetary counterparties) Corporate/government bonds Loans B2 Retail deposits L MM instruments Retail mortgage loans RML Bank bonds Reserves RES CB credit facilities DEP – CBDCb MM2 BB1 + BB2 RES + BAN – B3 – BB2 + CBDCb Central bank CB credit facilities Corporate/government/bank bonds 1 RES + BAN – B3 – BB2 Reserves + CBDCb Banknotes B3 + BB2 CBDC RES BAN – CBDCa CBDCa + CBDCb The analysis is performed under the assumption of a central bank operating through a corridor system. Explanatory notes: CBDCa – amount of banknotes substituted for by households’ CBDC holdings; CBDCb – amount of retail deposits at commercial banks substituted for by households’ CBDC holdings; RA1 (RA2) – real assets held by households (corporates/government); MM1 (MM2) – money market instruments issued by corporates/government (banks); B1/B2/B3 – amount of bonds (either issued by corporates or government) held by households/banks/central bank; BB1 (BB2) – amount of bonds issued by banks and held by household (central bank). In practice, however, some funding losses and gains and thereby some degree of deleveraging and/or leveraging are likely to happen as central bank credit leads to bank asset encumbrance. This, in turn, is costly to banks and may induce them to reduce their loans and bond holdings. To the extent that the shift in the structure of financial intermediation provokes higher (lower) liquidity, term and credit-risk premia on the funding for households and corporates/government, their capacity to hold real assets may decrease (increase). Central bank digital currencies 25 Annex C: The impact of CBDC on seigniorage Seigniorage represents income earned by a central bank from issuing (non-interest-bearing) banknotes. In a two-tier banking system, income from issuing money (banknotes and deposits at commercial banks) partly accrues to commercial banks, giving way to a broader notion of seigniorage. The design features of CBDC (described in section 2.2) determine how much of this broad seigniorage value accrues to commercial banks and to the central bank. If CBDC emerges as an attractive asset, seigniorage may move from commercial banks to the central bank, as agents substitute commercial bank deposits by CBDC. There are two channels through which broad seigniorage value may change due to CBDC. First, CBDC affects the overall value of the money issuing function to the extent that CBDC reduces operational costs (eg costs related to printing, storage and transportation of banknotes, and settlement costs) and, especially at the outset, entails significant fixed infrastructure costs (but very low marginal costs). Second, as an additional and possibly attractive asset, CBDC may serve as a substitute for other non-deposit financial assets (eg shares in money market mutual funds). This latter effect would increase money in circulation and thereby broaden the overall seigniorage base. Seigniorage accruing to the central bank depends on two key variables: the stock of currency in circulation and the difference in returns between central bank assets and currency liabilities. Introducing CBDC could change both factors. First, any CBDC-driven expansion of the balance sheet has a positive effect because most the funding cost equals the policy rate (ie the risk-free rate). Any asset that the central bank may buy from, lend to, or accept as collateral from its monetary counterparties should have an expected yield above the expected risk-free rate over the investment horizon. As a CBDC-driven expansion of the balance sheet entails a corresponding decline of retail deposits and money market instruments, such increased central bank seigniorage corresponds to decreased seigniorage income at banks and money market issuers. This effect may, however, be offset to some degree if CBDC were to lead to reduced demand for banknotes, which are non-interest bearing. And the impact would depend on the remuneration of CBDC: the higher the remuneration, the greater the reduction in seigniorage income from banknote circulation. These effects would produce gains and losses for central and commercial banks, as well as for nonbanks, which, in turn, could influence their financial robustness and hence have systemic financial stability consequences. For central banks, any significant reduction of seigniorage would constrain their ability to recapitalise following financial losses, in the absence of other sources of income. The persistence of low or even negative capital could put monetary policy and financial stability goals at risk. 26 Central bank digital currencies Annex D: Members of the working groups Committee on Payments and Market Infrastructures Chair Klaus Löber (European Central Bank) Reserve Bank of Australia David Emery National Bank of Belgium Filip Caron Central Bank of Brazil Daniel Gersten Reiss Bank of Canada Ben Fung European Central Bank Dirk Bullmann Bank of France Marion Chich Deutsche Bundesbank Heike Winter and Marcus Härtel Hong Kong Monetary Authority Nelson Chow Reserve Bank of India Supriyo Bhattacharjee Bank of Italy Michela Tocci and Giuseppe Galano Bank of Japan Shuji Kobayakawa Bank of Korea Dong Sup Kim Bank of Mexico Ángel Salazar Sotelo Netherlands Bank Kirsten van Driel Central Bank of the Russian Federation Maxim Grigoriev Saudi Arabian Monetary Authority Mohsen Al Zahrani Monetary Authority of Singapore Tze Hon Lau South African Reserve Bank Arif Ismail Sveriges Riksbank Björn Segendorf Swiss National Bank Marco Cecchini and Nino Landerer Bank of England Simon Scorer Board of Governors of the Federal Reserve System David Mills and Brendan Malone Federal Reserve Bank of New York Vanessa Lee Bank for International Settlements Paul Wong (Secretary) Morten Bech and Stijn Claessens Workstreams were led by Dirk Bullmann (European Central Bank), Shuji Kobayakawa (Bank of Japan), David Emery (Reserve Bank of Australia) and Brendan Malone (Board of Governors of the Federal Reserve System). Significant contributions were also made by Jiamin Lim (Reserve Bank of Australia); Hanna Halaburda (Bank of Canada); Thomas Leach (European Central Bank); Dion Reijnders (Netherlands Bank); Cordelia Kafetz (Bank of England); Jeff Marquardt and Sarah Wright (Board of Governors of the Federal Reserve System); Antoine Martin and Ray Fisher (Federal Reserve Bank of New York); and Ayse Sungur, Rebecca Chmielewski, Henry Holden, Rodney Garratt and Codruta Boar (Secretariat). Central bank digital currencies 27 Markets Committee Chair Aerdt Houben (Netherlands Bank) National Bank of Belgium Arnoud Stevens Bank of Canada Parnell Chu and Scott Hendry European Central Bank Jens Tapking and Christoph Ohlerich Deutsche Bundesbank Corinna Dietzen and Dorothee Hellmuth Hong Kong Monetary Authority Nelson Chow Reserve Bank of India Senthil Kumar Bank of Italy Tommaso Perez Bank of Korea Dong Sup Kim Netherlands Bank Steef Akerboom and Dion Reijnders Monetary Authority of Singapore Jeremy Hor Bank of Spain Covadonga Martín Alonso Sveriges Riksbank Per Åsberg Sommar Bank of England Cordelia Kafetz, Ben Dyson and Emily Clayton Board of Governors of the Federal Reserve System Laura Lipscomb and Heather Wiggins Federal Reserve Bank of New York Elizabeth Caviness Bank for International Settlements Andreas Schrimpf (Secretary) Stijn Claessens 28 Central bank digital currencies BIS Papers No 101 Proceeding with caution – a survey on central bank digital currency By Christian Barontini and Henry Holden Monetary and Economic Department January 2019 JEL classification: E42, E58, O33 Keywords: Central bank digital currencies, CBDC, digital innovation, money flower, cryptocurrencies, cryptoassets, financial inclusion The views expressed are those of the authors and not necessarily the views of the BIS. This publication is available on the BIS website (www.bis.org). © Bank for International Settlements 2019. All rights reserved. Brief excerpts may be reproduced or translated provided the source is stated. ISSN 1609-0381 (print) ISBN 978-92-9259-238-7 (print) ISSN 1682-7651 (online) ISBN 978-92-9259-239-4 (online) Contents Introduction ............................................................................................................................................... 1 Central bank digital currencies........................................................................................................... 1 Motivations for general purpose CBDCs: Sweden and Uruguay .......................................... 3 The survey .................................................................................................................................................. 6 Results .......................................................................................................................................................... 7 Conclusion ................................................................................................................................................12 Cryptocurrencies and other private digital tokens...................................................................14 References ................................................................................................................................................16 Annex 1: Central banks participating in the survey .................................................................17 Annex 2: Survey questions .................................................................................................................18 Previous volumes in this series ........................................................................................................ 20 BIS Papers No 101 i ii BIS Papers No 101 Proceeding with caution – a survey on central bank digital currency 1 A survey of central banks shows that a majority are collaboratively looking at the implications of a central bank digital currency. Although many have reached the stage of considering practical issues, central banks appear to be proceeding cautiously and few report plans to issue a digital currency in the short or medium term. Introduction Payments are changing at an accelerating pace. Users expect faster, easier payments anywhere and at any time, mirroring the digitalisation and convenience of other aspects of life (Bech et al (2017)). And, although paper-based payments like cheques and cash still play important roles, new technologies and market entrants are challenging the traditional bank-based payment systems (Jakobsen (2018)). In addition to changes in how payments are made, even the type of money used could be changing. Across the world, central banks are reportedly thinking about how new central bank digital currencies (CBDCs) could replace traditional money (CPMIMC (2018)). There is significant public interest in such a fundamental potential change, and this paper takes stock of central banks’ current work and thinking. It is based on a recent survey of central banks to which 63 responded 2 (representing jurisdictions covering close to 80% of the world population). The survey asked central banks about their current work on CBDCs, what motivates that work, and how likely their issuance of a CBDC is. The survey shows that, although a majority of central banks are researching CBDCs, this work is primarily conceptual and only a few intend to issue a CBDC in the short to medium term. Central bank digital currencies The 2018 report by the Committee on Payments and Market Infrastructures (CPMI) and the Markets Committee (MC) defines CBDCs as new variants of central bank money different from physical cash or central bank reserve/settlement accounts. Based on four key properties, the CPMI-MC report provides a taxonomy of money (“The money flower”) which delineates between two broad types of CBDC: general purpose and wholesale – with the former type coming in two varieties (Graph 1). The four key properties of money are: issuer (central bank or not); form (digital or physical); accessibility (widely or restricted); and technology. In terms of technology, 1 We thank Morten Bech and Paul Wong for valuable comments, Codruta Boar for excellent research assistance, Harish Natarajan and World Bank colleagues for help disseminating the survey, and Klaus Löber and members of the CPMI Working Group on Digital Innovations for comments on the questions asked. The views expressed in this article are those of the authors and do not necessarily reflect those of the BIS. 2 See complete list in Annex 1. BIS Papers No 101 1 the report distinguishes between money that is token- or account-based. In payment economics, a key difference between tokens and accounts is in their verification: a person receiving a token will verify that the token is genuine, whereas an intermediary verifies the identity of an account holder (Green (2008) and Kahn and Roberds (2009)). However, the definition of tokens varies considerable across scientific fields, and other reports distinguish between value- or account-based forms of CBDC (eg Sveriges Riksbank (2018) and Norges Bank (2018)). This paper uses the terms value- and token-based interchangeably. The money flower: a taxonomy of money Graph 1 The Venn diagram illustrates the four key properties of money: issuer (central bank or not); form (digital or physical); accessibility (widely or restricted); and technology (account-based or token-based). CB = central bank. Private digital tokens (general purpose) include cryptocurrencies, such as Bitcoin. For examples of how other forms of money may fit in the diagram, please refer to the source. Sources: CPMI-MC (2018); Bech and Garratt (2017). In sum, this paper discusses the three variants of CBDC highlighted by the greyshaded areas within the “money flower” above. The first is a “general purpose”, “account-based” variant, ie an account at the central bank for the general public. This would be widely available and primarily targeted at retail transactions (but also available for broader use). The second form is a “general purpose”, “token-based” variant, ie a type of “digital cash” issued by the central bank for the general public. This second variant would have similar availability and functions to the first, but would be distributed and transferred differently. The last form is a “wholesale”, “token- or value-based” variant, ie a restricted-access digital token for wholesale settlements (eg interbank payments, or securities settlement). Two general purpose CBDC projects, the e-Peso and e-Krona, and the motivations behind them, are discussed in detail in Box A. 2 BIS Papers No 101 Box A Motivations for general purpose CBDCs: Sweden and Uruguay Although in many parts of the world, the amount of cash in circulation has risen over the last decade, there are some countries that buck the trend (Bech et al (2018)). In this small club of jurisdictions, a few have considered general purpose CBDCs that would be a complement to cash. Sweden and Uruguay are notable not just for the advanced stage of their work but the amount of information their central banks have made publicly available about their respective projects. e-Krona Cash use in Sweden has declined for many years (Graph A). The country’s retailers have good reason to expect that the decline will continue and the cost of accepting cash will become prohibitive, so that it will no longer be accepted in the future (Sveriges Riksbank (2018)). Sveriges Riksbank payment survey As a percentage of respondents Who paid for their most recent purchase in cash? Graph A How often do you withdraw cash from an ATM or cash desk at a bank? Source: Sveriges Riksbank. In response to this decline, the Riksbank is working on an “e-Krona” project, beginning in early 2017 and publishing its second report in October 2018. The report noted that the use of cash continues to decline and that the state needs to have a role in the payment market. A means by which to do this is to have an electronic krona. The ekrona would be a complement to cash, as well as to current electronic payments (especially in a serious crisis where other electronic payments might not be available). Electronic payments beyond cards (specifically, a mobile payment system called “Swish”) have recently seen a significant increase in Sweden, but usage is markedly lower among the elderly (Graph B). The Riksbank notes that some in society, who may have access only to cash, including the elderly and other more vulnerable groups, may need a simpler, more user-friendly offering to avoid exclusion. An e-krona might be “value-based” (ie not an account). However, the current versions of distributed ledger technology (DLT) are considered to be too immature, although they are not ruled out for the future. The Riksbank envisages a “platform” where payment service providers (PSPs) of the e-Krona would connect and distribute the currency. Those PSPs could, the Riksbank thinks, use DLT in providing their services. A value-based approach would be compatible with the Riksbank’s legal mandate (the Sveriges Riksbank Act), but an account-based e-Krona would require the mandate to be adapted for clarity. An account-based e-Krona is not BIS Papers No 101 3 ruled out, but the Riksbank notes that coordination with other agencies would be necessary, and so dialogue should begin. The next stage will be a pilot programme for a prepaid value, non-interest bearing and traceable e-Krona. This will investigate a range of possible choices to better inform the decision whether to issue a full-scale e-Krona. Rapid increase in use of new payment solution in Sweden Sweden Means of payment % of GDP Graph B Means of payment for different age groups % of GDP Per cent Per cent Sources: Sveriges Riksbank; CPMI Red Book statistics. e-Peso The Central Bank of Uruguay has just completed a pilot programme on a general purpose CBDC. The pilot was part of a wider governmental financial inclusion programme, which began in 2011, aiming for greater access, labour market formalisation and payment system efficiency. Since these efforts began, the availability of ATMs and other cash dispensing mechanisms has grown enormously but cash withdrawals have plateaued (Graph C) and cash in circulation has fallen (Graph D). Uruguay Graph C POS Terminals ATMs 000s POS = point of sale. ATMs = automated teller machines. Cash withdrawals Number Mn Electronic payments UYU bn OEMP = other electronic means of payment. Source: Central Bank of Uruguay. 4 BIS Papers No 101 mn To respond to these changes and further its broader financial inclusion goals, the Central Bank of Uruguay began a pilot programme in November 2017 to issue, circulate and test an e-Peso. Unique digital banknotes in several denominations were issued for distribution to an “e-note manager platform”. The platform acted as registry of the ownership of the digital banknotes. DLT was not used. In total, 20 million e-Pesos were issued, of which 7 million were distributed by a third-party PSP, which held an equivalent value of pesos in a central bank account. Individual users and businesses, in electronic wallets, could hold a maximum of 30,000 e-Pesos (roughly USD 1,000) and 200,000 ePesos respectively. Transfers took place instantly and peer-to-peer, via mobile phones using either text messages or the e-Peso app. The Central Bank of Uruguay’s legal mandate was sufficient to issue the electronic e-Peso as a complement to physical cash. The pilot was deemed a success and closed in April 2018, after which all e-Pesos were cancelled. The programme is now in an evaluation phase and a number of questions are being considered, before a decision on further trials and potential issuance can be made. These include design specific challenges, eg how best to manage the stock of digital banknotes in different denominations as well as wider questions eg the level of anonymity the e-Peso would have, whether it would bear interest, the final role of the central bank and what the wider impact on businesses and the economy would be. M1 and its components in Uruguay As a percentage of GDP Graph D Source: Central Bank of Uruguay.  Sveriges Riksbank (2018).  Slides presented at the Conference on "Economics of Payments IX", hosted by the Bank for International Settlements and Committee on Payments and Market Infrastructures in Basel, Switzerland, 15–16 November 2018 (agenda). BIS Papers No 101 5 The survey Geographical coverage Some 63 central banks replied to the survey, of which 41 are located in emerging market economies (EMEs) and 22 in advanced economies (Graph 2). Together, the respondents represent close to 80% of the world’s population and over 90% of its economic output. Respondents to the survey1 Graph 2 1 The black circles represent the Cayman Islands, the Dominican Republic, the Dutch Caribbean, the euro area, Hong Kong SAR, Samoa, Singapore, the Solomon Islands and Tonga. “Advanced economies” and “Emerging market economies” as defined by the IMF World Economic Outlook country classification. The boundaries and names shown and the designations used in this map do not imply endorsement or acceptance by the BIS. Questionnaire The survey was conducted in latter part of 2018. It starts by asking central banks if they work on CBDCs or not and, if they do, it further inquiries about the type of CBDC and how advanced the work is. Motivations and current expectations for potentially issuing a CBDC are also queried, as well as whether central banks have legal authority to issue. The questions asked are included in Annex 2. Given the complexities involved, central banks also provided a wealth of supplementary qualitative explanations to their answers. This survey follows a similar (but smaller-scale and unpublished) survey conducted by the CPMI in 2017. Results from the 2017 survey are included where relevant. In addition to questions about CBDC, the survey also asked about “private digital tokens” and their use for payments. Private digital tokens encompass the wide variety of digital tokens not issued by central banks. The survey differentiated between socalled “cryptocurrencies” and other private digital tokens, with cryptocurrencies defined as decentralised tokens without an issuer or representing an underlying asset or liability. Central banks’ responses are discussed in Box B. 6 BIS Papers No 101 Results The survey finds that a wide variety of motivations is driving an increasing number of central banks to conduct conceptual research on CBDCs. However, only a few central banks have firm intentions to issue a CBDC within the next decade. Work underway Some 70% of respondents are currently (or will soon be) engaged in CBDC work, a slight increase compared to the 2017 survey (Graph 3, left-hand panel). Central banks currently not looking at CBDC are typically from smaller jurisdictions and/or face more pressing priorities. Some central banks indicate that they rely on research conducted by international organisations (in particular the BIS) or regional networks (eg CARICOM’s fintech Advisory Work Group). Of those that are engaged in work, over half cover both general purpose and wholesale CBDCs (Graph 3, right-hand panel), with about a third focusing only on general purpose and an eighth only on wholesale. Central bank CBDC work Share of respondents Graph 3 Engagement in CBDC work 1 Focus of work1 Share of respondents conducting work on CBDCs, 2018 survey. Source: Central bank survey on CBDCs. All central banks have begun their CBDC work with theoretical and conceptual research and are generally sharing their studies, with a view to developing a common understanding of this new field of study. At this point, half have moved on to experiments or more “hands-on” proof-of-concept activities to test new technologies (Graph 4, left-hand panel). This represents an increase of 15 percentage points over 2017 (Graph 4, right-hand panel). Many central banks in both advanced economies and EMEs are attempting to replicate wholesale payment systems using distributed ledger technology (eg projects Jasper, Ubin and Khokha (Bank of Canada (2018), Monetary Authority of Singapore (2018), South African Reserve Bank (2018)). BIS Papers No 101 7 Only five central banks have progressed to running pilot projects. The e-Peso project described in Box A is an example of a completed pilot. Importantly, despite the quantity of work underway, many of these proofs-of-concept or even pilot projects are only investigative in nature and do not imply plans to issue a CBDC. Type of CBDC work Share of respondents conducting work on CBDCs 2018 survey Graph 4 Experiments/proof-of-concept Source: Central bank survey on CBDCs. Central banks are also increasingly collaborating with each other to carry out proof-of-concept work on eg cross border payment and securities settlement arrangements. Collaborations include Project Stella by the ECB and the Bank of Japan (ECB-BoJ (2017)) as well as a joint project by the Bank of Canada (BoC), the Monetary Authority of Singapore (MAS) and the Bank of England (BoE) (BoC, MAS and BoE (2018)). Motivations The survey asked central banks about their motivations for potentially issuing a wholesale or a general purpose CBDC. Central banks chose from the same set of predefined factors for each type of CBDC, concerning payment safety and efficiency as well as other aspects of central banks’ mandates. The central banks ranked their relative importance on a four-point scale ranging from “not so important = 1” to “very important = 4” and supplemented their choice with comments. Looking across all respondents for both types of CBDC, payments safety and domestic efficiency are the most important motivating factors to central banks (Graph 5). Least important are, predictably, financial inclusion for wholesale CBDCs and, lesspredictably, cross-border payments efficiency, for general purpose CBDCs. To note, however, all rankings remain in a rather narrow range which suggest at this “investigative” stage the main motivation is to learn. However, as central banks progress, more differentiation in terms of motives might emerge. Outside the predefined choices, many central banks consider a range of other factors important as well. For general purpose CBDC, this broadly relates to issues around cash, either responding to dwindling use or discouraging it through supporting electronic innovations and payments. For wholesale CBDC, the other 8 BIS Papers No 101 factors are more diverse and overall, considered less important. They include better monitoring of transactions as well as safety and efficiency benefits for end users. Motivations for issuing a CBDC, ranked in order of importance Score1 Graph 5 General-purpose CBDCs Wholesale CBDCs 1 The score is calculated as an average of the options: “Not so important” (1), “Somewhat important” (2), “Important” (3) and “Very important” (4). Source: Central bank survey on CBDCs. Breaking respondents down by stage of economic development shows that, for general purpose CBDC, EMEs value domestic payments efficiency and financial inclusion most (Graph 6). On the other hand, cross-border payments efficiency is the least important. In contrast, for advanced economies, payments safety and financial stability are the primary motivators for potential issuance. Financial inclusion is clearly the least important factor. In qualitative commentary, EME central banks also note that supporting digitalisation, incorporating the informal economy and fighting financial crime, are key motivators for potentially issuing a CBDC. Some advanced economies are motivated by the prospect of a “less-cash” or even “cash-less” society (see Box A for a discussion of the e-Krona). BIS Papers No 101 9 Motivations for issuing general-purpose CBDCs, ranked in order of importance Score1 Advanced economies Graph 6 Emerging market economies 1 The score is calculated as an average of the options: “Not so important” (1), “Somewhat important” (2), “Important” (3) and “Very important” (4). Source: Central bank survey on CBDCs. For wholesale CBDCs, both advanced economies and EMEs consider payments safety and efficiency the most important motivating factors (Graph 7). However, for EMEs, the cross-border dimension is somewhat less important. All central banks (including EMEs) consider financial inclusion the least important factor for wholesale CBDCs. Motivations for issuing wholesale CBDCs, ranked in order of importance Score1 Advanced economies Graph 7 Emerging market economies 1 The score is calculated as an average of the options: “Not so important” (1), “Somewhat important” (2), “Important” (3) and “Very important” (4). Source: Central bank survey on CBDCs. 10 BIS Papers No 101 Outlook The survey asked central banks to describe the likelihood of their issuing each type of CBDC over the short (up to three years) and medium (up to six years) term. Central banks could choose from “very likely” to “very unlikely” on a five-point scale. In the short term, over 85% of central banks see themselves as either somewhat unlikely or very unlikely to issue any type of CBDC (Graph 8). No central banks are very likely to issue a wholesale CBDC in the short term, but two EME central banks are considering issuing a general purpose CBDC over the same horizon. Beyond the short term, an increased proportion of central banks consider the issuance of both types of CBDC to be possible. Nevertheless, a majority still consider this move at least somewhat or very unlikely. In the medium term, only one central bank reported that they see themselves as very likely to issue a wholesale CBDC. Overall, the likelihood of issuing both types of CBDC is somewhat similar, despite the perceived greater operational complexity and larger impact on the financial system of a general purpose CBDC (CPMI-MC (2018)). Likelihood of issuing a CBDC in the short and medium term1 Share of respondents Graph 8 General-purpose CBDC 1 Wholesale CBDC Short term: 1–3 years and medium term: 1–6 years. Source: Central bank survey on CBDCs. The 2017 survey also asked about the likelihood of issuing CBDC. However, the questionnaire did not differentiate between general purpose and wholesale CBDCs. Of the central banks that answered, half deemed issuance possible whereas the other half deemed it unlikely. At that time only one central bank was considering CBDC issuance to be likely but in the 2018 survey, it indicated that it is no longer pursuing any research. Legal authority A prerequisite for issuing a CBDC is that the central bank has the legal authority to do so. The survey asked central banks to indicate whether they have, or are in the process of acquiring, this authority. The same question was asked in the 2017 survey. BIS Papers No 101 11 Almost a quarter of central banks have, or will soon have, authority to issue a CBDC while a third do not, and about 40% remain unsure (Graph 9, left-hand panel). The high level of uncertainty is unsurprising, given that most central bank mandates predate not only cryptocurrencies but also many forms of electronic money. However, as central banks are studying all aspects of CBDC, the level of uncertainty has fallen compared with the 2017 survey (Graph 9, right-hand panel). The uncertainty does not differ materially by geography or a jurisdiction’s economic development. Legal authority to issue a CBDC Share of respondents 2018 survey 1 Graph 9 Over time There was no “laws are currently being changed to allow for it” option for the survey in 2017. Source: Central bank survey on CBDCs. Conclusion Most central banks are conducting research into CBDC. Many are progressing from conceptual work into experimentation and proofs-of-concept, including in cooperation with other central banks. Nonetheless, motivations for issuing a CBDC are largely idiosyncratic (eg falling availability of cash in a jurisdiction). This has meant that only a limited number of central banks are proceeding to the pilot stage with CBDCs, and even fewer see issuance of a CBDC as likely in the short or medium term. At this stage, most central banks appear to have clarified the challenges of launching a CBDC but they are not yet convinced that the benefits will outweigh the costs. Those that do see clear benefits are predominantly from EME jurisdictions. From survey responses, this seems to be because financial inclusion projects create a clear mandate for central bank action, and a lack of current infrastructure limits the disruption a CBDC could create while simultaneously encouraging the use of new technology. The trends identified in the survey are likely to continue. Different central banks will continue to move at different speeds. This creates a potential risk for spillover effects across borders (CPMI-MC (2018)). However, the evidence from this survey is that central banks are proceeding cautiously, and also that they are collaborating and 12 BIS Papers No 101 sharing the results of their work. Caution and collaboration will reduce the likelihood of unintended consequences. To meet the payment needs of the future, physical cash is unlikely to be the main answer. Yet, most people will have to wait to use a CBDC. However, central banks are working hard to make sure the wait is worth it. BIS Papers No 101 13 Box B Cryptocurrencies and other private digital tokens As well as questions on CBDC, the survey asked central banks about private digital tokens, encompassing the wide variety of digital tokens not issued by central banks. Decentralised digital tokens without an issuer that are not representative of any underlying asset or a liability are referred to as “cryptocurrencies”. The survey included questions on the use of cryptocurrencies for domestic and cross-border payments, their judgement on whether that use would rise or fall and the state of experimentation with other digital tokens by the private sector in their jurisdictions. Cryptocurrencies No central banks reported any significant or wider public use of cryptocurrencies for either domestic or cross-border payments in their jurisdictions (Graph E). Usage of cryptocurrencies is assessed to be either minimal (“trivial / no use”) or concentrated in niche groups for a large majority of the responding central banks. Answers were largely based on judgment, informed by industry, market and research sources although a few transaction monitoring programmes are reportedly in place. This is consistent with other research looking at payments made with cryptocurrencies (Graph G, right-hand panel), (Auer (2019)). Current use of cryptocurrencies for payments1 Share of respondents Domestic 1 Graph E Cross-border There were no responses for the options “Significant use” and “Wider public use”. Source: Central bank survey on CBDCs. Judgments about future usage are, unsurprisingly, difficult to make. Most central banks have not formed a firm view, especially in the case of cross-border payments (Graph F). Of those that could, the majority think use in payments will remain minor. Reasons for this judgment include low retail acceptance, compliance issues, better public understanding by the general public of the risks involved and, for some jurisdictions, outright bans. This is in line with other research that suggests the values and volumes of cryptocurrencies are influenced by regulators’ actions (Auer and Claessens (2018)). Some central banks reported that both the current and prospective use of cryptocurrencies seemed contained to such assets being used for investment purposes. 14 BIS Papers No 101 Anticipated use of cryptocurrencies for payments Share of respondents Graph F Domestic Cross-border Source: Central bank survey on CBDCs. Other private digital tokens A great deal of attention has been paid to the distributed ledger technology (DLT) underlying cryptocurrencies, with almost a quarter of respondents reporting that banks or non-banks are experimenting with or issuing private digital tokens as part of their payment services (Graph G, left-hand panel). The reported experiments are concentrated in advanced economies and remittance-receiving EMEs in Asia and are mostly at early stages. Projects reportedly focus on cross-border payments, consistent with domestic faster payments being available in the relevant jurisdictions. Some central banks note that initiatives are often akin to more traditional arrangements (eg e-money or correspondent banking) and may blur boundaries or give rise to definitional issues Are banks or non-banks experimenting with or issuing private digital tokens as part of their payment services? 2018 Survey Graph G Bitcoin payment transactions Per cent USD mn per month Sources: Auer (2019); Central bank survey on CBDCs. BIS Papers No 101 15 References Auer, R (2019): “The mechanics of decentralised trust in Bitcoin and the block chain”, BIS Working Papers, forthcoming. Auer, R and S Claessens (2018): “Regulating cryptocurrencies: assessing market reactions”, BIS Quarterly Review, September, pp 51–65. Bank of Canada (2018): Jasper phase 3 – securities settlement using distributed ledger technology, October. Bank of Canada, Bank of England and Monetary Authority of Singapore (2018): Crossborder interbank payments and settlements – emerging opportunities for digital transformation, November. Bech, M, Y Shimizu and P Wong (2017): “The quest for speed in payments”, BIS Quarterly Review, March, pp 57–68. Bech, M, and R Garratt (2017): “Central bank cryptocurrencies”, BIS Quarterly Review, September, pp 55–70. Bech, M, U Faruqui, F Ougaard and C Picillo (2018): “Payments are a-changin’ – but cash still rules”, BIS Quarterly Review, March, pp 67–80. Committee on Payments and Market Infrastructures (2015): Digital currencies, November. ——— (2017): Distributed ledger technology in payment, clearing and settlement: an analytical framework, February. ——— (2018): Cross-border retail payments, February. Committee on Payments and Market Infrastructures and Markets Committee (2018): Central bank digital currencies, March. European Central Bank and Bank of Japan (2018): Securities settlement systems: delivery-versus-payment in a distributed ledger environment, March. Financial Stability Board (2018): Crypto-asset markets, October. G20 Finance Ministers and Central Bank Governors Meetings (2018): Communiqué, March. Green, E, (2008): “Some challenges for research in payment systems” in A Haldane, S Millard and V Saporta (eds), The Future of Payment Systems, Routledge. Jakobsen, M (2018): “Payments are a-changin' but traditional means are still here”, Commentary on the CPMI “Red Book” statistics, December. Kahn, C and W Roberds (2009): “Why pay? An introduction to payments economics”, Journal of Financial Intermediation, vol 18, no 3, January, pp 1–23. Monetary Authority of Singapore (2017): Project Ubin phase 2 – re-imagining interbank real-time gross settlement system using distributed ledger technologies, November. Norges Bank (2018): Central bank digital currencies, Norges Bank Papers, no 1 2018, May. South African Reserve Bank (2018): Project Khokha – Exploring the use of distributed ledger technology for interbank payments settlement in South Africa, June. Sveriges Riksbank (2018): The Riksbank’s e-krona project – Report 2, October. 16 BIS Papers No 101 Annex 1: Central banks participating in the survey Some 63 central banks participated in the survey from the following jurisdictions: - Argentina Australia Azerbaijan Bangladesh Belgium Brazil Cambodia Canada Cape Verde Cayman Islands China Colombia Curaçao & Sint Maarten Cyprus Dominican Republic Ecuador Egypt Euro area (ECB) France Georgia Germany BIS Papers No 101 - Hong Kong SAR Hungary India Indonesia Iraq Israel Italy Jamaica Japan Jordan Kazakhstan Korea Kosovo Latvia Malaysia Montenegro Morocco Netherlands Nigeria Norway Pakistan Papua New Guinea - Philippines Russia Samoa Saudi Arabia Serbia Singapore Slovenia Solomon Islands South Africa Spain Sweden Switzerland Thailand Tonga Turkey United Kingdom United States Uruguay Vietnam Zambia 17 Annex 2: Survey questions 1. Has your central bank engaged, or will engage, in any kind of research, experiments or development work related to the development and use of CBDC? [Yes / No] 2. Is your work related to: - wholesale CBDC: general purpose CBDC both 3. What type of work is being, or will be, conducted? Please check all that apply. - research/ study experiments / proof-of-concept Development / pilot arrangement 4. How important are the following aspects to your motivations in issuing a: - General purpose CBDC Wholesale CBDC The following aspects were proposed: o o o o o o o financial stability monetary policy implementation financial inclusion payments efficiency (domestic) payments efficiency (cross-border) payments safety / robustness others (please specify below) For each: very important / important / somewhat important / not so important 5. How likely is it that your central bank will issue a CBDC in the: - General purpose CBDC Wholesale CBDC For both, two time horizons were proposed: o short term (within the next three years) o medium term (four to six years) For each: very likely / somewhat likely / possible / somewhat unlikely / very unlikely 6. Does your central bank have the legal authority to issue a CBDC? - Yes / no / uncertain / laws are currently being changed to allow for it 7. Please provide any other details about CBDC and the thoughts and work in your jurisdiction, including your key motivations. 8. For your jurisdiction, please tick "True" or "False" for the following statements: - 18 Domestically: o There is a real-time-gross-settlement system (RTGS) available o The RTGS system settles more than one currency o There is a faster payment system used for domestic retail payments o There is broad participation by eligible financial institutions in the faster payments system o There is a legal framework for e-money BIS Papers No 101 - o Non-banks are active in issuing e-money cross-border: o Payment mechanisms for cross-border e-commerce are widely available o There are exchange or capital controls that apply to cross-border retail payments 9. In your jurisdiction, how significant do you think consumer use of cryptocurrencies or crypto-assets for payments is? - For domestic payments For cross-border payments For each, the following options were proposed: o o o o o Significant use Wider public use Use by niche groups Trivial / no use Do not know 10. In your jurisdiction, do you think consumer use of cryptocurrencies or cryptoassets for payments is increasing or decreasing? - For domestic payments For cross-border payments For each, the following options were proposed: o o o o increasing staying the same decreasing Do not know 11. In your jurisdiction, are banks or non-banks experimenting or issuing private digital tokens as part of their payment services? [Yes / no / don’t know] BIS Papers No 101 19 Previous volumes in this series No Title Issue date BIS Papers No 100 Globalisation and deglobalisation December 2018 BIS Papers No 99 Central banks and debt: emerging risks to the effectiveness of monetary policy in Africa? October 2018 BIS Papers No 98 Low for long or turning point? July 2018 BIS Papers No 97 Financial spillovers, spillbacks, and the scope for international macroprudential policy coordination April 2018 BIS Papers No 96 The price, real and financial effects of exchange rates March 2018 BIS Papers No 95 Frontiers of macrofinancial linkages January 2018 BIS Papers No 94 Macroprudential frameworks, implementation and relationship with other policies December 2017 BIS Papers No 93 Building Resilience to Global Risks: Challenges for African Central Banks August 2017 BIS Papers No 92 Long-term issues for central banks August 2017 BIS Papers No 91 Financial systems and the real economy March 2017 BIS Papers No 90 Foreign exchange liquidity in the Americas March 2017 BIS Papers No 89 Inflation mechanisms, expectations and monetary policy November 2016 BIS Papers No 88 Expanding the boundaries of monetary policy in Asia and the Pacific October 2016 BIS Papers No 87 Challenges of low commodity prices for Africa September 2016 BIS Papers No 86 Macroprudential policy September 2016 BIS Papers No 85 A spare tire for capital markets: fostering corporate bond markets in Asia June 2016 BIS Papers No 84 Towards a “new normal” in financial markets? May 2016 BIS Papers No 83 What do new forms of finance mean for EM central banks? November 2015 BIS Papers No 82 Cross-border financial linkages: challenges for monetary policy and financial stability October 2015 BIS Papers No 81 Currency carry trades in Latin America April 2015 All volumes are available on the BIS website (www.bis.org). 20 BIS Papers No 101