The Bank for International Settlements (BIS) wrote an article for the Annual Review of Economics that laid out work done on Central Bank Digital Currencies (CBDCs) to date 1. The paper puts CBDCs in context of all payment methods and considers why and how CBDCs are being developed and the challenges central banks are grappling with to create a functional solution that delivers identified theoretical benefits while, at the same time, doing no harm. The paper reviews the research done so far and the extent to which this has shaped the work done, or not, and the gaps in research. This paper puts in one place an overview of journey so far and the direction of travel at the time of writing.
CBDCs are being developed in the context of an increasingly digitised economy and centrality of data. This brings with it concerns about competition, payment system integrity and privacy. The developments are paying attention to microeconomic (operational architecture, technologies, privacy) and macroeconomic considerations (financial systems, financial stability and monetary policy).
A significant change is that central banks no longer treat the current payment system as a benchmark. The focus has moved on from just looking at ‘reserves for all’ to include interactions with commercial banks, the effectiveness of monetary policy and the implications for financial stability of CBDCs.
The work done means central banks are increasingly confident that with targeted economic design they can achieve public policy goals while also limiting systemic risks. As a result, discussions now focus on potential enhancements to payment inclusion and payment efficiency, as well as policy objectives of competition, data privacy and the integrity of the payment system. Some central banks view CBDCs as countering the threat of ‘Big Tech’ activity in the payment arena.
Digital money, denominated in a national unit of account which is a liability on a central bank. CBDCs can be retail or wholesale, account- or token-based, using distributed ledger or conventional technology. Central banks are settled on versions of a two-tier structure for CBDCs in the monetary system. In the definition of CBDCs, the work on cross-border payment policy is ahead of academic research.
The history of CBDC initiatives is covered in the first chapter of the paper and is graphically laid out in graph 1 in the paper, ‘The monetary system with a retail CBDC’.
Alongside the digitalisation of economies is the increased centrality of data in the economic and monetary system, especially personal data. Access to this data may allow reduced information asymmetries, lower costs and new forms of money (based on ‘money as a memory’ 2 and automation through smart programmes). Network effects may allow new private players rapidly to dominate the monetary system, which is likely to reduce competition.
Changes in the payment system which have increased the focus on CBDCs include,
The rapid increase in Bitcoin and other cryptocurrencies. The negatives of these products are their association with volatility, speculation, criminal activity and that they are energy hungry.
Private sector issued stablecoins. From a central bank perspective, these are only as good as the governance behind them. If they are backed by conventional money, they are only an extension of conventional money. They could fragment the liquidity of the monetary system and detract from cash as a co-ordinating device.
Entry of Big Tech into payments. Their presence and use of big data could be a disruption of platform-based business models. Their network effect makes sector concentration more likely and create data silos, reducing competition. It would bring data privacy and governance issues and raises challenges of safety and integrity of payment systems against criminal activity.
The pandemic has accelerated the adoption of digital payments potentially bringing the need for CBDC options to be ready closer.
The motivation for CBDCs is not homogenous across central banks (see graph 5 in the paper, ‘Motivations for issuing CBDCs’). Central banks in Emerging Market and Developing Economies (EMDE) place a significantly higher value on financial inclusion than other central banks.
Work on retail CBDCs is more advanced if economies are more digitised (measured by mobile cellular phone subscriptions), the capacity of countries to innovate (measured by the output score of the World Intellectual Property Organisation) and the size of the informal economy as a percentage of GDP. For wholesale CBDCs the greater the financial development of the country, measured by their financial development index and trade openness, the greater the interest. The though is that wholesale CBDCs will allow more efficient clearing and settlement.
Finally, a motivation is the risk of cryptocurrencies, stablecoins and other countries CBDCs to monetary sovereignty, particularly if the first two of these are based on business empires of individual digital platforms rather than boundaries of legal jurisdiction. The paper then goes on to argue that these fears are overstated for three reasons.
Dollarisation did not happen because of technology but because of a lack of trust in the stable value of the local currency. The existence and ability to access foreign CBDCs, therefore, is not in itself the reason people may choose to use them.
If CBDCs are account-based linked to identity, then agreement will be needed between central banks for people to use foreign CBDCs and data will be available.
Central banks have little interest in destabilising a neighbour.
Addressing a potential challenge to usurp the US Dollars current role as the world’s reserve currency, the paper suggests this is unlikely. It is not the motivation of the e-CNY or the Digital Dollar projects. Reserve currencies are chosen because the country has deep capital markets, safe assets and hedging capabilities. There is trust in the long term value of the currency, the soundness of the legal and regulatory system and the currency is used in international trade for invoicing. Again, the existence of foreign CBDCs is not in itself the reason people may choose to use a CBDC as a reserve currency.
The goals are payment inclusion and efficiency, competition, data privacy and integrity.
Central banks are exploring different versions of a two-tier system working with the private sector, see graph 4 in the paper, ‘Retail CBDC architectures and central bank-private sector cooperation’. The three options described are:
Direct CBDC: central banks handle retail payments and run the entire system.
Hybrid CBDC: central banks record retail balances.
Intermediated CBDC: central banks record wholesale balances.
If the central bank opts for the intermediated model where it has no information on retail transactions, it must regulate the private sector so that wholesale holdings always add up to the sum of all the retail accounts. If they don’t the central bank would have to honour claims of which it has no direct record.
A central bank may have to set low fees to offer a competitive alternative exerting pressure on the margins of private companies.
Distributed Ledger Technology (DLT) is either permissionless, like Bitcoin, based on the ‘proof of work’ approach, or it is permissioned, a network of known and vetted validators who jointly augment a ledger.
Permissionless involves costly computing where every batch of transactions has to be accompanied by a proof that a substantial quantum of otherwise useless computations has been performed. This is not only inefficient, but it can lead to multiple equilibria (ie. “forks”). The paper concludes that, ‘For all these reasons, permissionless DLT based on proof of work is not a viable technology for CBDCs.’ A permissioned system offers operational resilience because it is based on traditional systems storing data multiple times in different physical locations. An important requirement for this to work is that validators will need to be incentivised so records can be trusted.
A data trail allows price discrimination and the ability to assess credit worthiness, both highly valuable to private sector organisations. Today private citizens do not ‘internalise the full cost of digital payments’. The report suggests CBDCs can trade off concerns about bank disintermediation against the social value of maintaining privacy with deposits, CBDCs and cash.
Privacy does not mean anonymity. With CBDCs it does mean greater privacy than consumer have today, but this will come at a cost and the need for public review and oversight. Options such as zero-knowledge proofs in a software-based CBDC system, application programming interfaces (APIs) in a CBDC system that constrict data exchange to only the necessary information for any given transaction and give users greater control over the data that they generate have all been considered. Ultimately, of course, whatever the technical design there will always be the concern or risk that public authorities would use CBDC systems as an instrument for state surveillance and control.
Research papers consider CBDCs in the context of monopolistic, perfect and imperfect competition with the asset side (loans) across all the papers being regarded as operating in a competitive market.
The existence of CBDCs means commercial banks must compete for funds (people may choose to hold money in CBDCs rather than bank deposits).
Deposits held at a bank carry a liquidity premium and therefore earn a low rate of interest for their owner. For a commercial bank these deposits are cheap and so are a preferred source of bank funds. Low interest rates mean that the opportunity cost of people holding money in deposits is high.
If CBDCs are interest bearing, commercial banks will have to respond in order to attract funds. The impact of this is that the cost of funds will increase for commercial banks adversely affecting aggregate lending and investments 3.
Academic literature does not consider the commercial banks funding their activities by increasing retained earning, increasing equity funding or issuing long-term bonds. This is because of the challenges of modelling balance sheets in a dynamic model. The paper suggests these different funding models could have a positive effect on the banking sector.
The risk with the existence of CBDCs is that the loss of confidence in one bank would not just impact that bank, but all. The systemic risk is greater and, in fact, if banks have increased their equity funding, banks may have a lower equity buffer to deal with the crisis.
There are a number of mitigations that need considering. Firstly, visibility of the start of a run might be better since data on movements would be visible to a central bank in a way deposit withdrawals are not. Secondly, there may actually be less damage to the economy. Central bank deposits are protected against forced liquidation because deposits at central banks are not callable. Nobody can force a central bank to liquidate long term projects. This may allow the consequences of a run to be managed to lower the impact.
When considering mitigations and scenarios, the paper suggests putting limits of CBDC holdings would distort payment efficiency in all states of nature where financial fragility is not an issue, ie. in most countries. It points out that protecting against ‘too big to fail’ and banking sector resilience, such as deposit guarantees, cost money and is a subsidy to banks.
CBDCs could provide information central banks can exploit for monetary policy uses, for example being able to target money transfers and monetary policy. Theoretically all firms, not just banks, could access central bank lending, zero bound limits could be set, not just on reserves, and central banks could move from setting inflation targets to price level targets.
However, this would blur the line between fiscal and monetary interventions, which could complicate the central banks’ role.
It could also amplify the international spill over of shocks. The paper argues that central banks adjusting the remuneration rate on CBDCs flexibly could be used to dampen spill overs. There are options such as limiting the use of non-resident holdings and setting fees for large or frequent cross border transactions that could also be used.
Cross border payments get a lot of attention because they are expensive, slow, lack traceability and transparency and are inaccessible to some people. The frictions are rooted in different domestic payment systems, for example different opening hours, technical standards and data requirements. Making large scale changes to the existing system are, therefore, hard. The advent of CBDCs offers the opportunity to start afresh.
To overcome legacy technology and process issues, cross border CBDC payment systems will need to be interoperable with consistent standards and co-ordinated CBDC designs.
A quarter of the 50 central banks working on CBDCs are working on cross border payments. The options being explored are compatible CBDCs, interlinked CBDC systems and multi-CBDC models, effectively a single CBDC system with jointly set rules and operating system where participants mutually recognise each other. This has to be a global effort.
This section ends pointing out that there is little academic research in this area and policy is ahead of it.
1 - BIS Working Paper 976 published November 2021. ‘CBDC: Motives, Economic Implications and the Research Frontier’. Raphael Auer, Jon Frost, Leonardo Gambacorta, Cyril Monnet, Tara Rice and Hyun Song Shin.
2 - The ability to keep full digital ledgers of transactions increasingly makes it feasible for agents to have a full record of their interactions with other agents in the past – or money as memory.
3 - Some research suggests that so long as CBDC interest rates are less than 1.49%, commercial bank could actually increase.
