Center for Financial Innovation and Stability

Notes from the Vault
Larry D. Wall
April 2019

If a cryptocurrency, such as bitcoin, were ever to become an important part of the global financial system, would we see the creation of cryptocurrency banks? Many in the crypto community take the view that a cryptocurrency-based financial system would eliminate the need for banks, especially those that keep reserves equal to only a fraction of their deposits. An alternative view is that there is no technical reason why cryptocurrency banks could not develop and, indeed, that economics would favor the creation of such banks.

If fractional reserve banks are likely to arise in a cryptocurrency economy, that could have a variety of important implications for the operation of the financial system. Perhaps the most important is that the total amount of funds denominated in the cryptocurrency that could be used for payments would depend in large part on the amount of cryptocurrency transactions deposits at banks. The level and fluctuations of such deposits could then have a material impact on the value of the cryptocurrency and the stability of the financial system.

This post analyzes the likelihood of and implications of cryptocurrency banking under the assumption that one or more cryptocurrencies will become a significant part of the global financial system.¹ I explain the basics of fractional reserve banking (FRB) and discuss the likelihood that FRB would arise in a cryptocurrency environment. I then discuss some of the implications if cryptocurrency banks were to arise.

Traditional fractional reserve banking
The traditional core of banking is simple: obtain funds from depositors and lend those funds to borrowers. Most bank deposit accounts have a short maturity, and a large fraction are in transactions accounts that are payable upon demand and are readily transferable to third parties. Bank deposits typically pay interest, either explicitly or implicitly in the form of services provided at reduced cost. Bank loans tend to be relatively short term, albeit with longer maturity than bank deposits. Banks earn a profit from receiving more interest on their loans than they pay on their deposits.

Bank balance sheets also contain two other items needed to facilitate their core business. First, on the funding side, banks maintain equity from their owners, as depositors demand some cushion to absorb unexpected loan losses.² Bank loans tend to be relatively low risk, and in most cases, the equity supplied by their owners should be sufficient to absorb all the loan losses protecting the depositors from loss. However, very few banks have sufficient equity to absorb all possible losses, and depositors are at some risk of loss.

Second, on the asset side, banks hold reserves in the form of currency and deposits at the Federal Reserve. Banks must be able to honor term deposit withdrawals at maturity and immediate requests for withdrawal on transactions accounts. Ordinarily liquidity is not a problem for a bank, as deposit inflows and loan repayments approximately equal the amount of deposit outflows. However, banks typically hold some reserves for those times when outflows exceed inflows. In general, a bank is expected to hold only a fraction of its transactions accounts as reserves. Banks that hold fewer reserves than the amount of their deposits are called fractional reserve banks, or FRBs.

As with equity and loan losses, the possibility exists that a bank will suffer net outflows in excess of its reserve account. Depending upon the circumstances, a bank may be able to obtain additional funds by offering to pay higher deposit rates or borrowing from other banks. However, such additional funding may not be available in all cases, especially if the bank experiencing the withdrawals also appears to have large loan losses.

Could fractional reserve cryptobanks emerge?
The paper by Satoshi Nakamoto (2008) that started bitcoin was motivated in part by a desire to eliminate the need for trusted third parties such as banks. Subsequently, much of the enthusiasm for bitcoin has come from people who desire a currency that is free from the control of the state or banks. Indeed, some cryptocurrency supporters view FRBs as being "based on dishonesty" and an "illusion" (for example, see here and here). One post argues that cryptocurrency is an "existential threat" to banks, while another post posits that cryptocurrency will "replace the banking system in the next decade." Yet others, such as a post by CNBC senior editor John Carney in 2013, have argued that there is nothing preventing the creation of a bitcoin-based fractional reserve bank, and MasterCard has reportedly sought a patent on cryptocurrency fractional reserve banking.³ Could fractional reserve cryptobanks emerge?

For the purposes of this post, let us assume that one or more cryptocurrencies become widely accepted as a means of payment.⁴ This necessarily implies that the usage has spread beyond existing cryptocurrency investors to a substantial fraction of the general public. If one were further to assume that cryptocurrencies displace banks, cryptocurrency would need to be used in high-value transactions between large institutions (financial and nonfinancial).

Assuming that the protocol underlying the cryptocurrency does not preclude fractional reserve banking, the key question is whether such a bank would be economically viable.⁵ The historical record is clear: fractional reserve banks have existed in a variety of different times and geographies. For some historical examples dating back to ancient Greece, see a paper by former Atlanta Fed economist Gerald P. Dwyer Jr. and Universidad Carlos III de Madrid professor Margarita Samartín (2009).

Along with paying interest, another historical service provided by banks was warehousing the money, which provided depositors with a safe, convenient place to store their money and a mechanism for making payments. These services were especially valuable when gold was used as money, as gold is heavy and costly to move. However, even paper money is bulky, difficult to use in large transactions, and needs to be stored in a safe place. The need for this warehousing when money comes in a physical form arguably does not apply to cryptocurrencies. A post by Jeffrey Tucker, director of content for the Foundation for Economic Education, observes that bitcoin is "weightless and takes no physical space." Thus, in principle, individuals have no need for a place to "warehouse" their cryptocurrency. However, Tucker acknowledges that cryptocurrency warehousing is widespread with the use of institutions such as Coinbase, which "seem like banks" in that they hold the cryptographic keys to accounts for safety and to provide a relatively convenient means of accessing their funds. However, he argues these institutions maintain a full-reserve policy to avoid the risk of losing their depositors to banks that are fully reserved.

Benefits of bank warehousing
Tucker’s comment about institutions that seem like banks raises an important issue with cryptocurrencies. Although cryptocurrencies require no storage space, they are vulnerable to being lost if an owner loses the cryptographic keys or they are stolen. Thus, many holders of cryptocurrency rely on firms to warehouse their cryptocurrency in "wallets" to provide greater security and friendlier user access. Even these firms are not completely secure, though, with one estimate of 2018 losses due to theft in cryptocurrency exchanges and wallets of at least $950 million.

If a cryptocurrency were to become widely used, its users will need a way to warehouse it that is both safe and convenient. Most people want their money kept safe until they need to make a purchase, and then they want convenient, immediate access. A bank specializing in cryptocurrency could help meet consumers’ needs by investing in cybersecurity to provide greater safety and developing user-friendly apps to make accessing deposits more convenient.

Benefits of bank lending
The above suggests that firms specializing in warehousing are likely to be important if cryptocurrency becomes an important part of the financial system. However, these warehouses need not be fractional reserve banks. As Tucker suggests, they could be banks that hold all of their deposits in the base money (which is cryptocurrency). Why would people hold risky deposits where the funds are invested in loans rather than in a safe, fully backed account?

Professor Lawrence H. White provides an answer in his testimony to a House Subcommittee. People hold risky deposits because FRBs use part of the interest they earn on loans to pay interest to depositors. This interest may be explicit payment and/or the provision of implicit interest in the form of reduced fees for its warehousing and payments services. One might argue that contemporary depositors are more risk averse than their predecessors, as suggested by Tucker’s comments on complete reserve backing. However, the evidence suggesting that cryptocurrency holders are extremely risk averse seems to be contradicted by the losses they have taken in existing wallets and in investing in initial coin offerings (see my recent post). Moreover, the historical record suggests that many people are not that risk averse.

As an alternative to bank deposits, individuals could lend directly to borrowers to earn interest, a possibility facilitated by the recent development of internet-based peer-to-peer lending facilities in the United States. However, most peer-to-peer lenders are no longer focused on peer-to-peer lending itself but rely on institutional funding, as my colleague W. Scott Frame observed in 2015. Moreover, a bank has several advantages over direct lending. First, banks specialize in making loans and have developed sophisticated ways of evaluating borrowers and collecting on delinquent loans. Second, loans are risky and some will default, with losses to the lender. Banks anticipate this outcome and charge a credit risk premium that exceeds expected losses on a loan. Banks also hold widely diversified portfolios such that the credit risk premium they earn on the good loans almost always exceeds the losses on their bad loans. Finally, the bank owners provide some equity financing, which protects depositors by taking a first-loss position should losses exceed earnings.

An additional benefit of relying on a fractional reserve bank relative to direct lending is that the FRB provides greater liquidity to interest-bearing accounts than would a peer-to-peer loan made by an individual. As White observes, an FRB can lend out most of its money knowing that withdrawals rarely exceed incoming deposits by a large amount. Thus, the bank need only hold a fraction of its deposits in the form of reserves to cover its ordinary needs. Moreover, should reserves prove inadequate, a bank can usually borrow from the interbank markets and institutional money markets to obtain additional funding.

Although banks’ specialization and diversification allow them to provide depositors with greater credit protection and liquidity than they could obtain on their own, FRB deposits are still risky. Banks may have problems obtaining sufficient funds to meet net withdrawals, and some banks have failed with losses in excess of the loss-absorbing cushion provided by equity holders.⁶

Some consequences of cryptocurrency FRB
The supply of any given cryptocurrency is under the control of its protocol. In the case of bitcoin, the protocol provides gradually increasing amounts up to a maximum supply of 21 million. Many other cryptocurrencies have similarly limited supplies, but others provide for a computer protocol to control the supply.⁷ Supporters of cryptocurrency point to the limits on supply (and stability of that supply) as likely to result in stable prices or even deflation (an increase in the currency’s value relative to goods and services). These supporters argue that this would provide a significant advantage over current monetary systems, which are prone toward inflation.

Fractional reserve banking in a particular cryptocurrency, however, would change the cryptocurrency monetary system in important ways. FRBs cannot create more of the underlying cryptocurrency (for example, the maximum amount of bitcoin would remain at 21 million). What FRBs can provide is a close substitute where the backing for the deposits consists of loans, reserves, and other assets (such as Treasury bills).⁸ Indeed, one could view FRB cryptocurrency deposits as a way in which the financial system would more efficiently employ a scarce resource by creating a close substitute. In this case, the scarce cryptocurrency is being used as reserves to support the growth of FRB deposits denominated in that cryptocurrency. The result would be an increase in the supply of transactions’ balances denominated in the cryptocurrency. This increase would also likely reduce the purchasing power of the cryptocurrency below the value it would have been absent FRBs.

Investors in that cryptocurrency may view the reduction in value due to FRBs as an unwelcome development. However, there are two important offsetting considerations. First, the maximum amount of cryptocurrency deposits that may be supplied through FRBs is a finite multiple of the amount of outstanding cryptocurrency. For example, if FRBs held reserves equal to 10 percent of deposits, the maximum amount of cryptocurrency created by all FRBs would be 10 times the outstanding amount of that cryptocurrency (such as 210 million bitcoin).⁹ Second, some people may decide to hold more FRB cryptocurrency deposits than they would have held of the underlying cryptocurrency, given the greater convenience of deposits and the interest received on their deposits.

Although the maximum amount of deposits in FRBs is a fixed multiple of the outstanding cryptocurrency given the reserve ratio, the amount of deposits actually created can vary considerably. This variation could occur due to normal changes in demand over the business cycle or due to changes in demand related to perceptions of the financial strength of FRBs. That is, FRB cryptocurrency deposits could be subject to runs (large-scale withdrawals) if the FRB is perceived to have taken losses in excess of its capital. Moreover, if depositors are running due to concerns about the bank’s solvency, the bank may also be unable to borrow money from other banks and/or investors.

Runs on an individual bank are not entirely bad as they can force the closure of a poorly managed bank that should be put out of operation. However, a run on a large fraction of the banking system, sometimes called a banking panic, can have severe adverse consequences for the real (nonfinancial) economy. As a result, governments in developed countries use prudential regulation and supervision to reduce bank risk taking. Governments have also provided their central banks with lender of last resort authority to make collateralized loans to commercial banks; governments also provide deposit insurance so that most retail customers do not have an incentive to run on their bank.

The combination of regulation, supervision, lender of last resort, and deposit insurance is not a complete solution to the problem of bank risk taking, illiquidity, and insolvency. The financial systems of many countries have come under severe stress despite the existence of these three protective actions, most notably in the fall of 2008 with the financial crisis. Additionally, government involvement in banking—especially the provision of lender of last resort and deposit insurance—reduces depositors’ incentives to monitor and discipline risk taking by their bank. Nevertheless, the elected representatives of almost all developed countries have determined that banking supervision, lender of last resort, and deposit insurance are, on net, beneficial for their banking systems.

A cryptocurrency FRB could likely be made subject to prudential supervision and regulation similar to the way cryptocurrency exchanges are subject to anti-money laundering rules. However, governments are unlikely to provide deposit insurance and cannot provide lender of last resort facilities. The problem with governments providing deposit insurance is that the credibility of insurance ultimately depends upon the government’s ability to cover all potential losses, a possibility that does not exist with a cryptocurrency. Moreover, as David Andolfatto observes in a post, there can be no lender of last resort because no central bank can issue more cryptocurrency.

Conclusion
Eliminating the need for banks is one of the motivations behind the creation of many cryptocurrencies, including bitcoin. However, should a cryptocurrency become an important part of the financial system, banks are likely to arise for some of the same reasons they exist in the current financial system. These cryptocurrency banks are likely to face the same risks as existing banks, but central banks and governments are likely to be in a weaker position to mitigate those risks.

References

Dwyer Jr, Gerald P., and Margarita Samartín. "Why Do Banks Promise to Pay Par on Demand?" Journal of Financial Stability 5, no. 2 (2009): 147–169.

Larry D. Wall is executive director of the Center for Financial Innovation and Stability at the Atlanta Fed. The author thanks Scott Frame for helpful comments. The view expressed here are the author’s and not necessarily those of the Federal Reserve Bank of Atlanta or the Federal Reserve System. If you wish to comment on this post, please email atl.nftv.mailbox@atl.frb.org..

_______________________________________

¹ Whether cryptocurrencies are likely to become a significant part of the financial system is an interesting topic but beyond the scope of this post. For a discussion of some of the issues raised by cryptocurrencies and their underlying technology, see here, here, and here.

² Virtually all U.S. banks have government-backed deposit insurance that lowers the risk to depositors, but means that the deposit insurer is taking on the risk that the value of a bank’s deposits will exceed the value of its loans.

³ Along these lines, there are companies such as BlockFi that take deposits and make loans in cryptocurrency, albeit it is not clear whether BlockFi’s deposits could be used as a transactions account.

⁴ As noted above, this assumption is not intended to say anything about the probability of one or more cryptocurrencies becoming widely accepted.

⁵ In order for a cryptocurrency to preclude fractional reserve banking, it would need to have the ability to censor (block) certain types of transactions. However, one of the widely cited virtues of existing cryptocurrencies is that they are censor resistant; the only requirement for completing a transaction is knowledge of the relevant public and private keys.

⁶ A post by Nic Carter notes another potential advantage of having banks for at least some cryptocurrencies. Many cryptocurrencies currently have limits on their transactions volume. A bank could help increase volume by providing an off-chain method of making payments.

⁷ Although the current protocol of a cryptocurrency determines the amount outstanding now and in the future, that protocol and its limits may be changed. See my post on the governance over changes in the blockchains that host cryptocurrencies.

⁸ In monetary economics terms, bitcoin would be outside money and deposits in a fractional reserve bitcoin bank would be inside money. See this Atlanta Fed post for a discussion of inside and outside money in the current financial system.

⁹ The maximum multiple that can be created is equal to one divided by the fraction of deposits held as reserves. See this post by Quickonomics.