Many of the central banks pursuing the goal of developing a Central Bank Digital Currency (CBDC), including early leaders, such as the Swedish Riksbank, are finding that designing a workable CBDC is proving extremely difficult. Many solutions would introduce significant changes, and thus considerable operational risks, into their respective banking systems; so much so, that the Bank for International Settlements (BIS) has called for an approach to develop a CBDC that uses ‘Minimally Invasive Technology’, arguing that “[g]iven society’s high dependence on a functioning and predictable monetary system, we call for a minimally invasive technology: one that offers consumers a digital complement to cash while preserving the two-tier monetary system and the important role of the private sector in it.”
The US Federal Reserve is also actively engaged in researching a CBDC, but as Chairman Jerome Powell has stated, the ultimate test will be whether there are “clear and tangible benefits that outweigh any costs and risks.”
But a ‘minimally invasive’ solution may be staring central bankers in the face.
In 1985, Nobel Prize winning economist James Tobin (1985) presented the case for what he called ‘segregated’ or ‘earmarked’ deposit liabilities – ‘earmarked’ for a specific purpose. He argued that one of those deposit types could be equivalent to central bank liabilities. Today, liabilities in such a segregated account would be a form of CBDC.
Assuming such a segregated account was ‘approved’ for a consumer’s use in an ‘Approved Financial Institution’ (AFI) of their choice, then when a consumer makes a deposit into such an account, the amount would become a central bank liability. Behind the scenes this would be reflected by the consumer’s bank debiting its reserve account at the central bank for the value of the deposit and the central bank recording the amount as its liability. Likewise, if a consumer makes a payment or a withdrawal from the account then the consumer’s bank would credit the central bank by the amount, and thus reduce the central bank’s overall liability.
In this model, the CBDC problem then becomes, at the highest level, a relatively simple accounting problem. No minting, issuing, or burning of digital currency ‘coins.’ No significant new technology required by the central bank, save for recording accounting entries and monitoring overall balances of these segregated accounts in the system. No significant new technologies to be developed by commercial banks. No need for the value of each receipt or payment to be reflected immediately on central bank books but merely on a regular basis, say daily. On a predetermined frequency, the net movements for each such account held by a bank would be aggregated and reported to the central bank, i.e., one number reflecting the day’s overall movement of digital cash for each approved institution. This has the added advantage that the central bank need not view each transaction across each account, merely the overall totals, therefore preserving consumer privacy. At any point in time, the total liability of the central bank would be the aggregate of the extant balances in all such segregated accounts in each bank that is ‘approved’ to operate these accounts.
Of course, there are operational problems with this basic model.
Even if such accounts paid no interest, in a crisis there could be a flood of money from private institutions into such accounts, potentially causing a bank run. The most obvious solution to that problem would be to set a limit – here called a Central Bank Liability Limit (CBLL) – for all approved accounts and to cap the central bank’s liability for each account at that limit. Such a system-wide limit could have other advantages, in that moving the limit up and down could encourage savings or spending, and thus could be used as a lever of monetary policy. Furthermore, if desired, a central bank could pay interest on the balance in such an account providing another lever to assist monetary policy.
Setting a CBLL however creates another problem – what about any balance above the CBLL? It is suggested here that any balance above the CBLL would become, as with other types of accounts, a liability of the consumer’s bank. These accounts would consist of two tiers: the balance in the lower Tier 1 being a liability of the central bank, and the balance in the higher Tier 2, if non-zero, a liability of the consumer’s bank. Here, such an account is termed a Composite Central Bank Account (CCBA) to reflect the composite nature of the balances in the two tiers. In accounting terms, only changes in those balances up to the CBBL need be reported to the central bank, ignoring any changes to Tier 2 balances purely above the CBLL and protecting consumers’ privacy.
In summary, under this model, the central bank’s liability for CBDC across the system would be the aggregate of all Tier 1 balances in all CCBA accounts in all AFIs. The maximum, or cap of the central bank’s liability for CBDC would be the sum of all CBLLs in all CCBA accounts in all AFIs. The aggerate would then be the current utilization of the overall CBLLs of CBDC across the system, and such information would be available in near real-time – in sharp contrast to coins and banknotes in circulation – allowing for finer monetary control.
It should be noted that such CCBA accounts would be able to take advantage of the technological innovations available to other types of ‘basic bank accounts,’ such as debit cards and internet and mobile banking. By making such an account low or no fee, like a ‘free checking’ account, this would assist in promoting another objective of a CBDC – improving financial inclusion.
However, for those citizens who do not wish, for whatever reason, to avail themselves of full banking facilities, they could access such a CCBA through a mobile phone wallet serviced by an approved AFI, provided that the identical CCBA structure, rules of operation, and central bank reporting requirements were adhered to. This again would improve financial inclusion, reduce – but not eliminate – the use of cash, and improve the overall efficiency of the financial system by promoting electronic payments and transfers.
One of the main advantages of such a ‘Minimally Invasive Technology’ model, is that it is based on current banking infrastructure and can scale easily to the volumes of payments required to support a CBDC, a large proportion of which will be redirected from existing payment accounts rather than new accounts.
Like any new concept, the devil is in the details, and implementing a CBDC will not merely be a technical exercise, but a legal one as well. There are many questions to be addressed before a CBDC is ready for prime time, in particular:
- Who should be ‘eligible’ to hold such CCBA accounts, e.g., would foreign citizens be permitted to hold them?
- How many such accounts should an eligible citizen be allowed to hold? It makes little sense to have limits on overall central bank liability if a consumer can disperse their wealth across an unlimited number of accounts.
- How can you improve privacy, even anonymity, of financial transactions in what should be a ‘cash-like’ account?
- How should the aggregate settlement of transactions across CCBA accounts between a central bank and AFIs be conducted?
- Which financial institutions would be ‘approved’ and what would be the approval process?
- How should central bank monetary policy be conducted?
- How should AFIs be remunerated for operating any type of CBDC system?
From a legal perspective, because this simpler CBDC model is based upon existing banking concepts and processes, it should not be overly difficult to adapt existing laws to accommodate the CCBAs described here. Furthermore, the responsibilities of the parties involved do not have to change much because: a central bank does not have to become a technology developer or provider; commercial banks will operate CCBAs as an integral part of, not parallel to, their existing systems and continue to develop innovations that support all their account types, including CCBAs; and consumers do not have to hold CBDC assets any differently than their other banking assets, and will have an integrated view of their financial affairs, including their holdings of CBDC.
A simpler CBDC, based on the model described by Tobin, could achieve the main policy objective of providing safer ‘central bank money’ to citizens without introducing the significant systemic operational risks incurred in other ‘parallel system’ models, such as that being trialed in China and Sweden.
Patrick McConnell has taught Strategic, Operational and Technology Risk Management at Macquarie University, Sydney and Trinity College, Dublin
This post follows on from his earlier blog post, “CBDC – How Dangerous is Programmability?”