BIS Working Papers: No 880 Rise of The Central Bank Digital Currencies: Drivers, Approaches and Technologies

BIS Working Papers
No 880
Rise of the central bank digital
currencies: drivers, approaches
and technologies
by Raphael Auer, Giulio Cornelli and Jon Frost
Monetary and Economic Department
August 2020
JEL classification: E58, G21.
Keywords: central bank digital currency, CBDC, central

banking, digital currency, digital money, distributed
ledger technology, blockchain.
BIS Working Papers No 880
Rise of the central bank digital currencies:
drivers, approaches and technologies∗
Raphael Auer,♦ Giulio Cornelli♦ and Jon Frost♦†
Abstract
Central bank digital currencies (CBDCs) are receiving more attention than ever before.
Yet the motivations for issuance vary across countries, as do the policy approaches
and technical designs. We investigate the economic and institutional drivers of CBDC
development and take stock of design efforts. We set out a comprehensive database
of technical approaches and policy stances on issuance, relying on central bank
speeches and technical reports. Most projects are found in digitised economies with
a high capacity for innovation. Work on retail CBDCs is more advanced where the
informal economy is larger. We next take stock of the technical design options. More
and more central banks are considering retail CBDC architectures in which the CBDC
is a direct cash-like claim on the central bank, but where the private sector handles
all customer-facing activity. We conclude with an in-depth description of three
distinct CBDC approaches by the central banks of China, Sweden and Canada.
Keywords: central bank digital currency, CBDC, central banking, digital currency,
digital money, distributed ledger technology, blockchain.
JEL classifications: E58, G21.
∗
The views expressed here are those of the authors and do not necessarily reflect those of the Bank
for International Settlements. For comments and input, we thank David Archer, Douglas Arner,
Codruta Boar, Harro Boven, Haiwei Cao, Stijn Claessens, Carl-Andreas Claussen, Sebastian Doerr,
Umar Faruqui, Leonardo Gambacorta, Philippe Haene, Henry Holden, Martin Hood, Linda Jeng, Taejin
Park, Mateo Piccolo, Martin Summer, Tara Rice, Amber Wadsworth, Peter Wierts, Hye-Rim Yoo,
Nouran Youssef, Dirk Zetzsche and participants at a BIS research meeting, an Arab Monetary Fund
Fintech Working Group meeting and a meeting of the OECD Experts Group on Finance and
Digitalisation. We thank Mu Changchun, Lyu Yuan, Scott Hendry, Francisco Rivadeneyra, Dinesh
Shah, Gabriela Guibourg, Martin Johansson, Hanna Armelius and Stig Johansson for conversations
regarding the CBDC approaches of the People’s Bank of China, Bank of Canada and Sveriges
Riksbank.
Bank for International Settlements (BIS), Cambridge Centre for Alternative Finance (CCAF).
♦ †
Rise of the CBDCs: drivers, approaches and technologies 1

1. Introduction
Over the centuries, wave after wave of new payment technologies has emerged to
meet societal demands. Coins, banknotes, cheques and credit cards were each
innovations in their own day (Giannini (2011)). Today, there is growing discussion of
a new payment technology: central bank digital currencies (CBDCs). As a digital
liability of the central bank, wholesale CBDCs could become a new instrument for
settlement between financial institutions, while retail (or general purpose) CBDCs
would be a central bank liability accessible to all. 1
Although the concept of a CBDC was proposed decades ago (ie Tobin (1987)),
attitudes about whether central banks should issue them have changed noticeably
over the past year. Initially, central banks focused on systemic implications that
warranted caution (Barontini and Holden (2019)). But over time, the need to respond
to the declining use of cash in some countries came to the fore, and a number of
central banks have warmed to the idea of issuing a CBDC. 2 A tipping point was the
announcement of Facebook’s Libra and the ensuing public sector response. As of late
2019, central banks representing a fifth of the world’s population reported that they
were likely to issue CBDCs very soon (Boar et al (2020)). Similarly, the share of central
banks (by number) that are likely to issue a retail CBDC over the medium term (in one
to six years) doubled in 2019, to 20%. Meanwhile, a full 80% of surveyed central banks
are engaging in research, experimentation or development of CBDCs. 3
During the Covid-19 pandemic, social distancing measures, public concerns that
cash may transmit the Covid-19 virus and new government-to-person payment
schemes have further sped up the shift toward digital payments, and may give a
further impetus to CBDC (see Auer et al (2020b)).
As a result, CBDCs have seized global attention and feature broadly in central
bank communications and public search interest (Graph 1). Still, no major jurisdiction
has decided to issue a retail CBDC, and many open questions remain. In the growing
literature on CBDCs, discussions centre on several fundamental aspects. One is how
central banks should create money and whether CBDCs are desirable in that context
(Keister and Sanches (2019), Jackson and Pennacchi (2019), Kim and Kang (2019),
Armelius et al (2020a)). Another area is the systemic implications of CBDCs and how
to cope with them (Brunnermeier and Niepelt (2019), Fernández-Villaverde et al
(2020), Kwon et al (2020), Carletti et al (2020)). There is also work on policy design
frameworks (Davoodalhosseini and Rivadeneyra (2020), Agur et al (2019), Allen et al
(2020)), their implications for cross-country payments (Milkau (2019)), implications for
the international role of currencies (Ferrari et al (2020) and legal aspects of their
issuance (Hess (2020), Duque (2020), Nabilou (2020), Belke and Beretta (2019)).
1
“General purpose” and “retail” are used interchangeably to refer to CBDCs that individuals and non-
financial firms could access. For an overview and relevant definitions, see Bech and Garratt (2017)
and CPMI and Markets Committee (2018).
2
Neither electronic money nor the discussion on the central bank’s role in providing it directly to the
people is new (ie Tobin (1987)). In the context of CBDCs, Broadbent (2016), Liikanen (2016), Mersch
(2016), Wilkins (2016), Menon (2016), Skingsley (2016) and Nakaso (2016) were among high-level
policymakers who argued early on that the idea should be taken seriously.
3
Also in the private sector, the potential design and implications thereof are garnering substantial
attention. See PwC (2019) and World Economic Forum (2020).

CBDCs: the next hype or the future of payments? Graph 1
Timing of speeches and reports on CBDC1 Google search interest over time2
Number of speeches Search interest by year, index
1
12-month moving sum of the count of central bankers’ speeches resulting from a case-insensitive search for any of the following
words/phrases: CBDC; central bank digital currency; digital currency and digital money. 2 12-week moving average of worldwide search
interest. The data has been normalised to the 12-week moving average peak of each series. The search was run on search terms “Bitcoin” and
“Facebook Libra” and topic “Central Bank Digital Currency”. Data accessed on 16 July 2020.
Sources: Central bankers’ speeches; central banks’ websites; Google Trends; authors’ calculations.
Finally, the technology of retail CBDCs and how they relate to private sector
proposals is hotly contested (see Auer and Böhme (2020), Klein et al (2020), Clark and
Mihailov (2019), Brunnermeier et al (2019) and Vives (2019)). Much less contentious
is the issuance of wholesale CBDCs (see Bech et al (2020) and Pfister (2020)).
Amidst intense discussion in the research and policy spheres (BIS (2020)), and
early development efforts, this study analyses the economic and institutional drivers
of CBDC projects, thus shedding light on ultimate motivations. A next step is to
understand the policy approaches and technical design of the various projects, and
to look for commonalities and differences across countries.
The questions this paper aims to answer are: what are the economic and
institutional drivers for issuing CBDCs? How do central banks approach the issues?
What are the technical solutions sought? To answer these questions, we first develop
a novel CBDC project index based on central bank research and development (R&D)
projects. We then empirically investigate common factors in countries that are
investigating and piloting CBDCs. We find that higher mobile phone usage (a measure
of an economy’s overall digitisation) and higher innovation capacity are positively
associated with the likelihood that a country is currently researching or developing a
CBDC. Retail CBDCs are more likely where there is a larger informal economy, and
wholesale CBDCs are more advanced in economies that have higher financial
development.
Next, we look at four attributes of CBDC technical designs, following the
taxonomy of Auer and Böhme (2020). We show that a rising number of central banks
are considering “Hybrid” or “Intermediated” architectures where the CBDC is a cash-
like direct claim on the central bank, but the private sector manages customer-facing
activity. Only a small number of jurisdictions is considering designs in which the
central bank takes on an important operational role in the customer-facing side of

payments. None of the central bank reports favour a design with indirect claims on
the central bank (referred to as an “Indirect” or “Synthetic” CBDC architecture).
Whereas many central banks are considering multiple technological options
simultaneously, current proofs-of-concept tend to be based on distributed ledger
technology (DLT) rather than a conventional technological infrastructure.
Nevertheless, access frameworks tend to be based on account identification rather
than allowing for token-based fully anonymous access. Most CBDC projects have a
domestic focus. We examine how the features of these CBDCs fit with each other and
the unique economic structures and preferences of their populations.
Finally, we show that the circumstances of each jurisdiction also matter for the
policy approach taken to researching and developing a CBDC. Based on public reports
and in-depth interviews with the experts in respective central banks, we describe
three advanced approaches: the Chinese Digital Currency Electronic Payments
(DC/EP), the Swedish e-krona and the Bank of Canada’s CBDC contingency plan. While
these projects are each tailored to their national context, there are lessons for other
jurisdictions.
The rest of the paper is organised as follows. Section 2 describes current CBDC
research and development efforts. Section 3 conducts empirical analysis on the
drivers of CBDC projects. Section 4 discusses policy approaches and technical design.
It gives a stocktake of four attributes of CBDC projects and relates these to economic
indicators. Section 5 discusses the three selected projects. Section 6 concludes with
policy implications and avenues for future research. The data collection process is
described in Annex A. Annex B gives a tabular overview of CBDC projects. There is a
separate online annex with further empirical results.
2. Taking stock of CBDC research and development efforts
Central banks around the world have been researching the concept and design of
digital currencies for several years. As early as 2014, the Central Bank of Ecuador
launched a project called “Dinero electrónico” (electronic money) to allow individuals
to make mobile payments through a central bank-operated system (Valencia (2015)).
Yet the system failed to attract a significant number of users, and was discontinued
in 2016 (White (2018)).
Concurrently, with the growing popularity of Bitcoin and distributed ledger
technology (DLT), a number of central banks have started internal projects to better
understand DLT and its potential application to currencies. In the Netherlands, the
Netherlands Bank (“De Nederlandsche Bank”, DNB) did internal experiments starting
in 2015 with a DLT-based coin called the Dukaton (DNB (2018)). This was named after
the dukaat, a golden coin used at the time of Dutch independence from Spain in the
16th century. The Bank of England, Monetary Authority of Singapore, Bank of Canada
and others ran similar internal experiments around this time. They generally
concluded that DLT was not yet mature enough for use in major central bank payment
systems (see Bank of England (2017)). In March 2016, the Deputy Governor of the
Bank of England gave thoughts on the wide-reaching implications of CBDCs

(Broadbent (2016)) – the first of many subsequent policymaker speeches that have
grappled with CBDCs and their implications for central banks and societies. 4
From 2016 onward, a number of central banks launched research projects on
digital currencies for specific purposes. The Bank of Canada launched Project Jasper
(named after Jasper National Park in Alberta) in early 2016, and published a first
report about the work the next year (Bank of Canada (2017)). 5 This project initially
focused on DLT for the settlement of high-value interbank payments. The Monetary
Authority of Singapore launched its own Project Ubin (named after the island Pulau
Ubin) at the Singapore FinTech Festival in November 2016 (MAS (2016)). This, too,
focused on interbank payments, and specifically on a tokenised form of the Singapore
dollar on DLT. 6 The Hong Kong Monetary Authority (HKMA) launched Project
LionRock (named after a mountain in Hong Kong) in January 2017. The European
Central Bank (ECB) and the Bank of Japan launched the first known example of
cooperation between two central banks on CBDCs in 2017 with Project Stella, focused
on cross-border payments (ECB-BoJ (2017)). 7 Monetary authorities of Saudi Arabia
and the United Arab Emirates, and of Hong Kong SAR and Thailand, also announced
cross-border work on wholesale CBDCs (SAMA-UAECB (2019); BoT-HKMA (2020)). 8
The first publicly announced work on retail CBDCs was conducted by the Swedish
Riksbank (Sveriges Riksbank (2017)). In Sweden, cash use has been declining in recent
years, and the Riksbank has initiated a societal discussion on access to a central bank
payments instrument for the general public. Over time, this “e-krona” project has
been further developed. In February 2020, the Riksbank announced it would conduct
a pilot project with Accenture aimed at developing a proposal for a technical solution
for an e-krona (Sveriges Riksbank (2020); see Section 4). While announced somewhat
later, perhaps the most advanced CBDC project at present is that of the People’s Bank
of China (PBC). Known as the Digital Currency Electronic Payment (DC/EP), this CBDC
is now being piloted in four cities in China (see Section 4). DC/EP will be a cash-like
liability of the PBC available to the general public – and to foreign visitors of China –
through account-based interfaces.
Meanwhile, the Bank of Canada has announced that is does not currently see a
case for a retail CBDC, but that it is conducting work on retail CBDC as a contingency
plan in case cash use suddenly declines or a private digital currency is widely adopted
(see Section 4). The Eastern Caribbean Central Bank (ECCB) has launched a pilot called
DXCD (ECCB (2019)), and the Central Bank of the Bahamas has launched a pilot called
the Sand Dollar (CBB (2019)). Graph 2 gives a timeline of these and other projects.
4
There were a number of earlier speeches on private digital currencies. For instance, in December
2013, the Governor of the Central Bank of Mauritius discussed digital currencies in a speech on the
development of the local financial sector (Bheeninck (2013)). In 2015, the Deputy Governor of the
Reserve Bank of India discussed private digital currencies in a speech (Gandhi (2015))
5
In addition, the Bank launched a formal research agenda on electronic money and payments in 2013.
6
The Bank of Canada, Bank of England and MAS subsequently worked in collaborative research with
banks on challenges in cross-border payments and settlement and how various initiatives – including
wholesale CBDC – could help. See Bank of Canada-MAS-Bank of England-HSBC (2018).
7
Project Stella is so named because the flags of both the European Union (EU) and Japan depict stars.
The EU musters 12 stars. The Japanese flag depicts the sun – the star in our solar system.
8
The project by the Bank of Thailand and the HKMA is titled Inthanon-LionRock, after the tallest
mountain in Thailand (Doi Inthanon) and the iconic Lion Rock in Hong Kong. SAMA-UAECB Project
“Aber” refers to an Arabic word for a cross-road or moving from one place to another.

CBDC projects have proliferated since 2016 Graph 2
Sources: Central bankers’ speeches; central banks’ websites.
As of mid-July 2020, at least 36 central banks have published retail or wholesale

CBDC work (Graph 3). At least three countries (Ecuador, Ukraine and Uruguay) have
completed a retail CBDC pilot. Six retail CBDC pilots are ongoing: in the Bahamas,
Cambodia (Bomakara (2019)), China, the Eastern Caribbean Currency Union,
CBDC projects status Graph 3
BS = The Bahamas; ECCB = Eastern Caribbean central bank; HK = Hong Kong SAR; SG = Singapore.
The use of this map does not constitute, and should not be construed as constituting, an expression of a position by the BIS regarding the
legal status of, or sovereignty of any territory or its authorities, to the delimitation of international frontiers and boundaries and/or to the
name and designation of any territory, city or area.
Source: central banks’ websites.

Speeches on CBDCs have turned more positive since late 2018
Number of speeches Graph 4
Search on keywords “CBDC”, “digital currency” and “digital money”. The classification is based on the authors’ judgment. The score takes a
value of –1 if the speech stance was clearly negative or in case it was explicitly said that there was no specific plan at present to issue digital
currencies. It takes a value of +1 if the speech stance was clearly positive or a project/pilot was launched or was in the pipeline. Other speeches
(not displayed) have been classified as neutral.
Sources: centralbanking.com; Central bankers’ speeches; central banks’ websites; authors’ calculations.
Korea (Bank of Korea (2020)) and Sweden). 9 Meanwhile, 18 central banks have
published research on retail CBDCs (eg Harahap et al (2017), Burgos and Batavia
(2018), Kiselev (2019) and Bank of Japan (2020)), and another 13 have announced
research or development work on a wholesale CBDC.
In parallel, a growing number of central bank governors and board members
have made public speeches about CBDCs. In 2017 and 2018, many of these had a
negative or dismissive stance, particularly toward retail CBDCs. Since late 2018, the
number of positive mentions of retail and wholesale CBDCs in speeches has risen, and
in fact there have now been more speeches with a positive than a negative stance
(Graph 4).
The motivations for such work differ across jurisdictions. Based on a survey of
central banks in the BIS Committee on Payments and Market Infrastructures (CPMI)
in late 2019, Boar et al (2020) show that in advanced economies (AEs), central banks
are researching CBDCs to promote safety and robustness, or domestic payments
efficiency (Graph 5). Financial stability concerns may also be an important driver of
research and development work. Especially in emerging market economies (EMEs),
financial inclusion is an important motivation.
9
Additionally, the Marshall Islands, which currently use the US dollar as legal tender and do not have
a monetary authority, have launched the SOV project, a digital currency proposed by private
developers. We do not include this project in our database. See IMF (2018) for a critical discussion.

Motivations for issuing a CBDC
Average importance Graph 5
General purpose (retail) CBDC Wholesale CBDC
1 = not so important; 2 = somewhat important; 3 = important; and 4 = very important.
Sources: CPMI survey of central banks; Boar et al (2020).
Recently, the Covid-19 pandemic may have accelerated work on CBDCs in some
jurisdictions. For instance, in the United States, early versions of Congress bills on
fiscal stimulus included references to a “digital dollar” as a means of quickly executing
government-to-person payments, as an alternative to credit transfers and slow and
costly cheques (Brett (2020)). In parallel, the Federal Reserve has continued its
ongoing research on retail CBDCs (Brainard (2020a, b)). In the Netherlands, the central
bank has emphasised that the pandemic underscores the need for a backup to private
money (DNB (2020)). In China, pilot testing for the new CBDC is coinciding with a
phasing out of pandemic-related mobility restrictions. In Sweden, testing of the e-
krona project continues even amidst central bank crisis management measures.
It should be noted that retail payment behaviours show great inertia. For
example, Brown et al (2020) find that an exogenous introduction of more convenient
payment methods led only to a moderate average reduction in the cash share of
payments. Arifovic et al (2017) show with experimental evidence how fees influence
the behaviour of buyers and sellers, and ultimately the take-up of a new payment
method. However, when behaviours change, they often do so quite persistently. In
the same manner, changed payment behaviours caused by the Covid-19 crisis, such
as a greater use of digital payments, could have far-reaching effects in the future.
3. The cross-country drivers of CBDC development
Several global developments – including the digitalisation of commerce, the rise of

private digital currencies and concerns that cash may transmit the Covid-19 virus –
have recently driven increased interest in CBDCs. Yet the economic and institutional
motivations for issuance vary according to the country.

In this section, we seek to explain the cross-country dimension of CBDC research
and development based on economic and institutional drivers. Specifically, we want
to find commonalities in why central banks choose to embark on – or step up – CBDC
efforts in some countries more than in others, using cross section regressions. This
will also help us to understand how they design CBDC projects.
We thus build a CBDC project index, and then seek to explain its development
across countries and over time.
A novel CBDC database
We start by generating a novel global index measuring the central bank’s progress
toward the development of a retail or wholesale CBDC: the CBDC project index
(CBDCPI). This index captures publicly announced work by the central bank on CBDC
projects. This is equal to 0 when there is no announced project, 1 in the case of public
research studies, 2 in the case of an ongoing or completed pilot and (a so-far
hypothetical) 3 for a live CBDC. We construct one sub-index each for retail and
wholesale CBDC projects, and an overall index equal to the maximum of the two sub-
indices. The information was collected through desk research and with the help of
contacts at several individual central banks. 10
Separately, we have calculated two different indicators:
• A central bank speech score, which reflects discussion of CBDCs in speeches
by central bank management, including the stance (whether positive or
negative). This is set to –1 if the authors judge that the speech stance was
clearly negative or in the case it was explicitly said that there was no specific
plan at present to issue a CBDC. It takes a value of 0 in the case of a neutral
stance. Finally, it takes a value of +1 if the speech stance was clearly positive
or if a CBDC project/pilot was launched or was in the pipeline. For each
country, the aggregate score is a simple average of the individual speech
stances.
• A search interest index corresponding to Google Trends or (for China) Baidu

searches in each country on “CBDC” and related terms by the general
public. 11 This score is equal to the simple average of the search intensity of
the individual keywords for the period January 2013–April 2020. The score
reflects both interest by citizens in the idea of a CBDC, and how widely known
any central bank plans to introduce a CBDC are with the public.
For each of the indicators described above, we replace country-level missing

observations with zeros. This choice is consistent with the absence of a project
(research or pilot), a neutral stance towards the development of a CBDC (speech
score) or a lack of public interest (as captured by the search intensity score).
10
The list of projects is broadly consistent with other stocktakes, such as Kiff et al (2020) and Atlantic
Council (2020). We only take account of official central bank communications, not press articles.
11
For China, we use the Baidu index for keywords “Central Bank Digital Currency” and “DC/EP”. Baidu
index data directly measure the daily number of searches. To compare this with Google Trends data,
we re-index the data to a peak of 100. Next, we reduce the frequency, taking the monthly maximum.
Finally, we calculate a simple average of the two series for the period 2013–current. See Annex A for
a detailed explanation.

These indicators – made available with this paper – can help to gauge the project
work on CBDCs in specific countries and to compare it with communications by
central banks and public interest. The CBDC project index and both the speech and
search interest scores display substantial variance in the cross section. Naturally, the
three variables are correlated with one another, as central bank board members often
use speeches to broadcast project work, and public search interest may be higher
where central banks have communicated that they are working on a project (see
pairwise correlations in Table 1).
Pairwise correlations Table 1
CBDC project index1 Central bank Search interest index

(overall) speech score2 (Google/Baidu)3
CBDC project index1 1.00
(overall)
Central bank 0.52*** 1.00
speech score2
Search interest index 0.40*** 0.20*** 1.00
(Google/Baidu)3
*** denotes significance at the 1% level.
1
The project stance is equal to 0 when there is no known work on retail or wholesale CBDC, 1 in the case of research output, and 2 in the
case of an active or completed retail or wholesale CBDC pilot. 2 Search on keywords “CBDC”, “digital currency” and “digital money”. The
classification is based on the authors’ judgment. The score takes a value of –1 if the speech stance was clearly negative or in case it was
explicitly said that there was no specific plan at present to issue digital currencies. It takes a value of +1 if the speech stance was clearly
positive or a project/pilot was launched or was in the pipeline. Other speeches (not displayed) have been classified as neutral. Normalised
and winsorised at the 5% level. 3 Data have been normalised.
Sources: Baidu; central banks’ websites; BIS Central Bankers’ Speeches; Google Trends; authors’ calculations.
There are necessarily caveats to these measures. For instance, many central banks
have not publicly released reports on their ongoing CBDC projects. Some central
banks (eg PBC) have quite advanced projects, but have given relatively few speeches
on their plans. In some jurisdictions, Google (or Baidu) is not widely used for internet
searches. Still, the index can provide a comparable yardstick to assess changes across
countries and over time. Moreover, this can provide a useful complement to the
anonymised responses of central banks through official surveys. In the next section,
we try to explain the cross-country heterogeneity in the CBDC project index.
Examining the cross-country drivers of CBDC projects
In this section we investigate the drivers of the CBDC project index. To complement
central bank surveys and official motivations, we look at “revealed policy preferences”,
ie the economic and institutional factors that are associated with central banks’ actual
work on overall, retail or wholesale CBDCs. Our cross section estimations use an
ordered probit approach (McKelvey and Zavoina (1975)), and take the form of:
𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃(𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝑖𝑖 = 0,1,2,3|𝑥𝑥𝑖𝑖 ) = 𝐹𝐹(𝛼𝛼 + 𝛽𝛽𝑥𝑥𝑖𝑖 + 𝜀𝜀𝑖𝑖 )
where 𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃(𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝑖𝑖 = 0,1,2,3|𝑥𝑥𝑖𝑖 ) is the probability that the CBDC project index
(overall, or for retail or wholesale projects) in jurisdiction i equals 0 (no project), 1
(research), 2 (pilot) or 3 (live CBDC), F() is the functional form of ordered probit, Xi is
one or more variables from a vector of potential drivers, 𝛼𝛼 and 𝛽𝛽 are estimated
coefficients and εi is an error term.

Some of the potential drivers of CBDC development can be related to factors
affecting a country’s technological capability to develop and deploy a CBDC. Focusing
on indicators from reliable sources that are available for a wide cross section of
countries, we include in our analysis the following indicators:
- Digital infrastructure: jurisdictions with greater mobile phone use (mobile
cellular subscriptions per 100 people) or internet use (fixed line broadband
subscriptions per 100 people) may have a more developed infrastructure for
the central bank to develop CBDCs. Data on both come from the World Bank.
- Innovation capacity: jurisdictions with a higher innovation score overall,
and hence the ingenuity and R&D potential to support central banks in
designing a new CBDC ecosystem, may be more likely to see CBDCs. Data
come from the World Intellectual Property Organization (WIPO) Global
Innovation Index, which aggregates measures in the political environment,
education, infrastructure and business sophistication (WIPO (2018)). To look
at the innovation capacity of the central bank, itself, we have a dummy for
countries that have in place or plan to institute a retail fast payment system
(FPS). Data for this come from Bech and Boar (2019).
- Institutional quality: jurisdictions with higher government effectiveness
may be more likely to launch CBDC projects. Data come from the World Bank.
Conversely, central banks in jurisdictions with a large informal (“shadow”)
economy may have greater interest in creating a data trail for transactions,
and thus promoting use of a digital currency. Estimates of the size of the
informal economy come from Medina and Schneider (2019).
On the other side, countries may differ in their perceived demand for a CBDC. To
proxy these factors, we include the following indicators:
- Development and financial inclusion: countries that are more developed,
as measured by GDP per capita, may see a higher demand for new digital
payment methods. At the same time, all else equal, jurisdictions with lower
access to transaction accounts may see a greater need for retail CBDCs as
a financial inclusion policy. Data come from the World Bank Findex. 12
Meanwhile, jurisdictions with higher financial development may have
greater demands on innovative solutions for wholesale settlement; data for
this are available from Svirydzenka (2016).
- Public interest in CBDCs: where the public has more internet searches for
CBDCs and related topics, this may signal either that they are more aware of
the topic of CBDCs in general, or the plans of their own domestic central
bank in this area. Either way, a positive association can be expected. Data are
from Google and Baidu, as discussed above.
- Cross-border transactions: while most CBDCs serve a domestic purpose,
one could expect that especially internationally oriented CBDCs (eg projects
for cross-border interbank settlement or migrant remittances) may be more
likely in more internationally integrated economies. Trade openness (the
sum of imports and exports over GDP) can proxy for cross-border demand
for new payment options for goods and services. Remittance flows (inflows
12
We have also looked at various measures of cash use, such as small-denomination banknotes to GDP.
These are available from the CPMI Red Book statistics, but unfortunately for only 18 jurisdictions.

and outflows divided by GDP) gauge the economic importance of migrants’
remittances. Again, both series come from the World Bank.
Descriptive statistics1 Table 2
Standard
Observations Mean Min Max
deviation
Dependent variables
Overall CBDC project index 175 0.31 0.66 0 2
Retail CBDC project index 175 0.22 0.53 0 2
Wholesale CBDC project index 175 0.13 0.48 0 2
Independent variables
Mobile cellular subscriptions (per
100 people) 169 109.24 39.54 12.60 320.55
Broadband subscriptions (fixed line,
per 100 people) 167 13.60 13.38 0 47.16
Innovation output score (WIPO) 118 29.67 12.69 7.90 67.13
Fast payment system dummy 175 0.37 0.48 0 1
Government effectiveness 175 0.08 0.99 -2.24 2.19
Informal economy (% of GDP) 122 26.08 11.62 5.43 55.78
GDP per capita (USD) 168 16,652 21,423 301 110,344
Account ownership (% age 15+) 135 60.39 27.96 6.45 99.96
Financial development index2 158 0.36 0.22 0.06 0.93
Search interest index
(Google/Baidu)3 175 0.11 1.13 –0.34 8.18
Remittances to GDP
4
110 5.89 7.86 0.19 41.18
Trade openness 5
134 80.05 48.87 0 345.69
Central bankers' speech stance
index7 175 0.02 0.47 –0.13 1.68
1
For all the independent variables, average over the period 2013–19, subject to data availability.
2
Svirydzenka (2016). 3
Data have been
normalised. 4 Sum of inflows and outflows. 5 Sum of imports and exports divided by the country GDP. Data for 2018. 6
Data for 2014;
for EA, latest observation available (2010). 7 Normalised and winsorised at the 5% level.
Sources: Bech at al (2020); Medina and Schneider (2019); Svirydzenka (2016); WIPO (2018); IMF, World Economic Outlook; World Bank,
Remittance Prices Worldwide, remittanceprices.worldbank.org; World Bank; Baidu; central banks’ websites; BIS Central Bankers’ Speeches;
Datastream; Google Trends; authors’ calculations.
Table 2 gives descriptive statistics for our sample. For our CBDCPI, we have 175
observations. This includes a number of jurisdictions that are part of a currency union.
In these cases, we count jurisdictions only if they have a central bank that could in
theory develop a CBDC; currency unions without national central banks are
considered one observation, with all independent variables calculated as weighted
averages according to 2018 GDP. 13 Mobile cellular subscriptions range from 13 per
13
For this reason, the 19 euro area members are included individually, plus an additional observation
for the ECB, given the ECB’s research work (project score of 1). For individual euro area members,
values of the CBDC project index are 1 or 0 depending on work by national central banks. On the
other hand, the eight members of the Eastern Caribbean Currency Union (ECCU) are aggregated to
one observation, with a project score of 2 given the ECCB pilot. Empirical results are robust to
dropping the individual euro area member countries from the sample. See Annex A for data
construction.

100 people (North Korea) to 321 (Macao), and GDP per capita ranges from USD 281
(Burundi) to USD 110,343 (Luxembourg). For some key variables (eg the innovation
output score, estimates of the informal economy, account ownership and
remittances) coverage is lower, but generally still well above 100 jurisdictions.
Table 3 displays our univariate regression results. We can confirm that the CBDC
project index is strongly associated with higher mobile and internet use, a higher
innovation capacity, an existing or planned FPS and greater government
effectiveness. It is also higher in jurisdictions with higher search interest for CBDCs.
Somewhat against our expectations, there is a negative association with the informal
economy in these univariate estimations; as we will see below using a multivariate
approach, this relates to the correlation of this variable with mobile use and other
positively associated covariates. Further when it comes to those factors potentially
affecting the demand for CBDC, we find CBDC projects to be more advanced where
there is higher GDP per capita, financial development and search interest. Higher
account ownership is associated with more advanced CBDC project work, while
remittances are negatively correlated. As a simple robustness check, we have
performed the same analysis with ordinary least squares (OLS). Results (available in
an online annex) are broadly consistent. Similarly, the univariate results are very
similar for the retail and wholesale indices separately.
Of course, these simple regression coefficients need to be interpreted with great
care as many of the regressors are collinear (see online annex). More advanced
economies tend to be more digitised, more innovative and to feature more effective
goverments and smaller informal economies. Moreover, isolating individual drivers is
complicated by the fact that sample size for some indicators is more limited, thus not
allowing us to include all possible regressors at the same time.
To better control for multiple country characteristics, Table 4 displays
multivariate ordered probit regression results for the overal CBDC project index, and
for retail and wholesale CBDCs. Here, we confirm that overall projects are more likely
where there is greater use of mobile phones and greater innovation capacity. For
example, a one standard deviation increase in mobile phone subscriptions is
associated with a 55–63% higher probability of moving from no work to research, or
from research to a pilot (assuming mean levels of all independent variables). 14 A one
standard deviation increase in the innovation output score is associated with a 55%
higher probability. In this case, controlling for mobile use and other positively
associated covariates, we also find a significant association with the size of the
informal economy and financial development for the overall project index. We do not
find a significant link with trade openess.
14
There are necessarily caveats to this simple calculation given the non-linear nature of the ordered
probit and correlation between the independent variables. For a discussion of interpretation issues
in logits, probits and other non-linear probability models, see Breen et al (2018).

Univariate ordered probit regressions on overall CBDC project index1 Table 3
Digital infrastructure
Mobile cellular subscriptions 0.010***
(per 100 people) (0.004)
Broadband subscriptions 0.042***
(fixed line, per 100 people) (0.008)
Innovation capacity
Innovation output score 0.047***
(WIPO) (0.009)
Fast payment system (FPS) 0.882***
dummy
(0.221)
Institutional characteristics
Government effectiveness 0.674***
(0.118)
Informal economy –0.03***
(% of GDP) (0.013)
Development and financial
inclusion
Ln(GDP per capita) 0.439***
(0.092)
Account ownership 0.023***
(% age 15+) (0.005)
Financial development 3.414***
index2 (0.552)
Public interest in CBDCs
Search interest index 0.432***
(Google/Baidu)3 (0.098)
Cross-border transactions
Remittances4 to GDP –0.157**
(0.068)
Trade openness5 0.001
(0.003)
Number of observations 169 167 118 175 175 122 168 135 158 175 110 134
Pseudo R2 0.057 0.126 0.129 0.074 0.145 0.058 0.119 0.131 0.215 0.105 0.113 0.001
Robust standard errors in parentheses; ***/**/* denotes results that are significant at the 1/5/10% level.
1
For all the independent variables, average over the period 2013–19, subject to data availability. 2 Svirydzenka (2016). 3
Data have been
normalised. 4 Sum of inflows and outflows. 5 Sum of imports and exports divided by the country GDP. Data for 2018.
Sources: Bech et al (2020); WIPO (2018); Medina and Schneider (2019); Svirydzenka (2016); IMF, World Economic Outlook; World Bank,
Remittance Prices Worldwide, remittanceprices.worldbank.org; World Bank; Baidu; central banks’ websites; Datastream; Google Trends;
authors’ calculations.

Multivariate ordered probit regressions on CBDC project indices1 Table 4
Wholesale CBDC project

Overall CBDC project index Retail CBDC project index
index
Mobile cellular subscriptions 0.013** 0.015*** 0.011** 0.022**
(per 100 people) (0.005) (0.006) (0.005) (0.010)
Innovation output score 0.045*** 0.067*** 0.082***
(WIPO) (0.010) (0.017) (0.019)
Informal economy 0.027* 0.033* 0.042*** -0.009
(% of GDP) (0.015) (0.018) (0.016) (0.026)
Financial development 3.909*** 3.303*** 4.287***
Index2 (0.867) (0.775) (1.299)
Trade openness3 -0.003 -0.016** 0.004* -0.001
(0.004) (0.007) (0.003) (0.004)
Number of observations 118 105 110 100 132 105
Pseudo R 2
0.167 0.241 0.144 0.244 0.263 0.352
Robust standard errors in parentheses; ***/**/* denotes results that are significant at the 1/5/10% level. Constants are not reported.
1
For all the independent variables, average over the period 2013–2019, subject to data availability. 2
Svirydzenka (2016). 3
Sum of
imports and exports divided by the country GDP. Data for 2018.
Sources: Medina and Schneider (2019); WIPO (2018); IMF, World Economic Outlook; World Bank; central banks’ websites; Datastream; authors’
calculations.
Retail CBDCs, also appear to be more advanced in jurisdictions with high

innovation capacity and where the informal economy is larger, all else equal. A one
standard deviation increase in the size of the informal economy is associated with a
38–49% probability of moving one unit up in the CBDC project index, all else equal.
This result, obtained only when controlling for other factors, could relate to a desire
by authorities to have a data trail for transactions, as discussed above.
Wholesale CBDCs are positively correlated with financial development, which
could reflect the focus of such projects on increasing the efficiency of wholesale
settlement. 15 In the more parsimonious specification, there is a link with trade
openness. As many wholesale projects focus on the cross-border dimension, this link
is also intuitive.
These results, as well, are robust to using OLS (in online annex).
4. Policy approaches and technical design
We have thus far established that CBDCs are more likely to be under research and
development in jurisdictions with high mobile use, innovation capacity and search
interest for CBDCs, with some differences across retail and wholesale CBDCs. We have
also noted that CBDC projects differ starkly across countries, both in their economic
and institutional motivations, policy approach and their technical design.
15
The innovation output score is not included, given high correlation with financial development (81%).

In what follows, we focus only on the 30 retail CBDC projects in our sample (see
Annex C). We explore four key technological attributes of retail CBDC projects, and
the economic and institutional factors correlated with their use.
A stocktaking framework: the CBDC Pyramid
Approaches to CBDC design are heterogeneous across countries, requiring us to distil

the main design choices and the dimensions along which national approaches differ.
One way to classify design approaches is the “CBDC Pyramid” (see Auer and Böhme
(2020) and Graph 6). This approach starts from the consumer needs that a retail CBDC
could address, identifies associated technical design trade-offs, and then derives the
design choices. The scheme of design choices forms a hierarchy in which the lower,
initial layers represent design decisions that feed into subsequent, higher-level
decisions. To reflect this hierarchy, the choices are displayed as a pyramid.
The CBDC pyramid Graph 6
The CBDC pyramid maps consumer needs onto the associated design choices for the central bank. The left-hand side of the CBDC pyramid
sets out the consumer needs and associated features that would make a CBDC useful. The pyramid’s right-hand side lays out the associated
trade-off – forming a hierarchy in which the lower layers represent design choices that feed into subsequent, higher-level decisions.
Source: Auer and Böhme (2020).
The first and foundational design choice is the architecture, ie which operational
role the central bank and private intermediaries take on in a CBDC. Intermediaries can
run into technical difficulties or solvency issues. A CBDC should be safe from such
outages. Yet payment intermediaries offer valuable services to consumers, which are
needed to ensure the same level of convenience, innovation and efficiency as in
today’s payments. The architecture needs to balance these two considerations.

We augment Auer and Böhme (2020) by classifying various proposals for CBDC
design into four distinct CBDC architectures. These differ in the structure of legal
claims and the record kept by the central bank. They are:
• Direct CBDC – a payment system operated by the central bank, which offers
retail services. A CBDC is a direct claim on the central bank. The central bank
maintains the ledger of all transactions and executes retail payments.
• Hybrid CBDC – an intermediate solution that runs on two engines.
Intermediaries handle retail payments, but the CBDC is a direct claim on the
central bank, which also keeps a central ledger of all transactions and
operates a backup technical infrastructure allowing it to restart the payment
system if intermediaries fail.
• Intermediated CBDC – an architecture similar to Hybrid CBDC, but in which
the central bank maintains only a wholesale ledger, rather than a central
ledger of all retail transactions. Again, the CBDC is a claim on the central bank
and private intermediaries execute payments. For the purposes of this paper,
this will be considered alongside the Hybrid model in our stocktake.
In addition to these three generally recognised general purpose CBDC architectures,
another approach is the indirect provision of retail CBDC via financial intermediaries.
We note that, as this does not allow the consumer to directly access central bank
money, not all central banks recognise this architecture as a general purpose CBDC. 16
• Indirect or Synthetic CBDC – a payment system operated by
intermediaries that resemble narrow payment banks. Consumers have
claims on these intermediaries, which operate all retail payments. These
intermediaries need to fully back all liabilities to retail clients with claims
on the central bank. 17
The second technical design choice regards the infrastructure. A CBDC must be
secure from outages at the central bank. The infrastructure can be based on a
conventional centralised database or instead on DLT. These technologies differ in
their efficiency and degree of protection from single points of failure. DLT often aims
to replace trust in intermediaries with trust in an underlying technology. Calle and
Eidan (2020) describe some of these proofs-of-concept in detail. Also noteworthy is
that all central banks experimenting with DLT use permissioned variants, where
operators can decide who is admitted to the network. No central bank report
examined in this study has ventured to rely on permissionless DLT, as used for Bitcoin
and many other private cryptocurrencies. 18
The third choice concerns how consumers can access the CBDC. Account-based
CBDCs are tied to an identity scheme, which can serve as the basis for well-functioning
payments with good law enforcement. Yet access is likely to be difficult for one core
target group: the unbanked and individuals who rely on cash. There may be
challenges to match the qualities of cash as an inclusive, crisis-proof and anonymous
means of payment (Pichler et al (2019)). An alternative is to base access on so-called
16
For example, Sveriges Riksbank (2020), Bank of England (2020), and for the case of Canada, Dinesh
et al (2020) only consider architectures featuring direct claims on the central bank.
17
See Adrian and Griffoli-Mancini (2019) and Kumhof and Noone (2019).
18
See Auer (2019) for a discussion of the inefficient economics of permissionless models and Ali and
Narula (2020) for a specific analysis of permissioned and permissionless DLT in the context of CBDCs.

digital tokens. 19 This allows for value-based payment options, for example pre-paid
CBDC banknotes that can be exchanged both physically and digitally. Yet this also
brings new risks of illicit activity and counterfeiting.
Closely tied to the domestic access framework is the fourth design choice, the
use of CBDC for cross-border payments, which relates to the retail and wholesale
interlinkages in a CBDC’s design and its accessibility for residents vs non-residents.
Token-based domestic access would naturally be open to anyone, including non-
residents. But central banks may allow for use by non-residents.
Graph 7 classifies the attributes of ongoing retail CBDC projects. Among the retail
CBDC projects in our sample, we find a wide variety of approaches to architecture,
infrastructure, access and cross-border (retail or wholesale) interlinkages. On
architecture, we find four central banks considering a Direct model (often to enhance
financial inclusion). Seven are considering the Hybrid or Intermediated options (in
some cases alongside a Direct option), and a larger group has not yet specified the
architecture. No report examined in this study indicates that a central bank is pursuing
the Indirect/Synthetic architecture.
Attributes of retail CBDC projects

Number of retail CBDC projects investigating each design option Graph 7
Architecture Infrastructure Access Interlinkages
Interm. = Intermediated; Ind. = Indirect; Synth. = Synthetic; Undec. = undecided/unspecified or multiple options under consideration; DLT =
distributed ledger technology; Conv. = Conventional; Nation’l = national use; Internat’l = international use.
Sources: central banks’ websites; authors’ calculations.
Regarding infrastructure, we find seven central banks running their prototypes

on DLT, three with conventional technology and one considering both (Shah et al
(2020)). Yet these infrastructure choices are often for first proofs-of-concept or pilots.
Only time will tell if the same choices are made for large-scale designs. Among access
19
Importantly, this definition of token versus accounts must not be confused with the one used in the
field of computer science. Rather, it follows Kahn and Roberds (2009). As put by Kahn (2016), the
distinction between accounts and tokens are the identification requirements: “In a token‐based
system, the thing that must be identified for the payee to be satisfied with the validity of the payment
is the “thing” being transferred – “is this thing counterfeit or legitimate?” In an account‐based system,
however, the identification is of the customer – “Is this person who she says she is? Does she really
have an account with us?”

methods, account-based access appears to be the most common to date, with six
central banks clearly leaning toward account-based and a further two looking at both
account and token-based access. Finally, while most of the projects in our sample are
focused on domestic use, several of them – by the ECB, the central banks of France,
Spain and the Netherlands, and the ECCB – are by construction focused on cross-
border use among the members of a multi-country currency area.
The drivers of technological designs
Central banks choose these different attributes of CBDCs in line with the unique needs
of their jurisdictions, but there may nonetheless be common features across
countries. 20 In this light, we have also performed simple probit regressions for three
of the four attributes (Table 5; moreover, a figure in the online annex represents this
graphically). Indicators were chosen based on the statistical significance of differences
between the projects; the top three are presented in each case. Because of the lack
of variation in our sample regarding cross-border interlinkages, these differences are
not presented.
For the architecture, one may expect that the Direct or Hybrid and Intermediated
options would be more likely in economies that are less developed or have less
financial inclusion. In fact, we find the opposite: likely due to the influence of the
Nordic countries, Canada and China, we find that it is higher-income jurisdictions,
with greater account access and government effectiveness, that are more likely to
choose a Direct or Hybrid and Intermediated architecture – at least in research work
to date. Less developed countries have generally not specified their chosen
architecture.
Regarding infrastructure, we expect that DLT – originally designed to substitute for
trusted intermediaries – is more attractive in jurisdictions in which authorities are
perceived to be less effective. This difference is not statistically significant. Some
central banks do note explicitly that DLT presents no fundamental advantages when
using a centralised issuance system (NBU (2019)). On the other hand, countries
researching or piloting DLT are more dependent on inward remittance flows than
countries researching or piloting conventional architectures.
Regarding account-based vs token-based access, we find that countries looking
at tokens have higher public search interest, but also lower remittance inflows. These
differences are only significant at the 10% level.
20
We note that central banks often consider multiple design options, to retain full flexibility when it
comes to achieving the wide range of objectives such as privacy, monetary policy, inclusion, or
financial stability.

Univariate probit on retail CBDC project features1 Table 5
Architecture Infrastructure Access

1: Direct and Hybrid or
1: DLT infrastructure 1: Token-based
Intermediated
0: Indirect / unspecified 0: Conventional / unspecified 0: Account-based / unspecified
Government effectiveness 0.599**

(0.299)
Ln(GDP per capita) 0.600**
(0.249)
Account ownership (% age 15+) 0.033**
(0.015)
Remittance inflows to GDP 0.205**
(0.096)
Informal economy 0.02
(% of GDP) (0.023)
Trade openness2 0.006
(0.007)
Search interest index 0.157
(Google/Baidu)3 (0.131)
Remittance outflows to GDP -1.873**
(0.828)
Innovation output score -0.008
(WIPO) (0.022)
Number of observations 31 31 27 28 27 31 31 26 28
Pseudo R2 0.103 0.150 0.148 0.118 0.027 0.0191 0.046 0.186 0.005
Robust standard errors in parentheses; ** denotes results that are significant at the 5% level. Constants are not reported.
1
For all the independent variables, average over the period 2013–19, subject to data availability. 2
Sum of imports and exports divided by
the country GDP. Data for 2018. 3 Data have been normalised.
Sources: WIPO (2018); Medina and Schneider (2019); IMF, World Economic Outlook; World Bank, Remittance Prices Worldwide,
remittanceprices.worldbank.org; World Bank; Baidu; central banks’ websites; Datastream; Google Trends; authors’ calculations.
5. Approaches for CBDC design: three examples
In the above analysis, we have thus far conducted a study of the drivers of CBDC
interest and how the economic and institutional characteristics of each jurisdiction
shape technological CBDC design choices. The circumstances of each jurisdiction also
matter for the policy approach taken to researching and developing a CBDC. We
believe that there can be great value in central banks learning from one another’s
approaches – an activity that is promoted by international policy discussions (BIS
(2020)) and the publications of central banks. To complement this work, we describe
three unique CBDC approaches, one each in Asia, North America and Europe.
To make the discussion of design choices more concrete, we showcase three
prominent retail CBDC projects, namely the People’s Bank of China’s Digital Currency

Electronic Payment (DC/EP) project; the Swedish Riksbank’s e-krona; and the work by
the Bank of Canada on a CBDC as a contingency plan.
PBC: the DC/EP project (pilot stage)
Among all current CBDC projects, the one by the People’s Bank of China (PBC) is at
the most advanced stage. CBDC development efforts in China go back to at least
2014. In late 2019, the PBC announced it would conduct a pilot study for a retail CBDC,
the Digital Currency and Electronic Payment (DC/EP) project. On 20 April 2020, a PBC
spokesperson confirmed that pilot testing was under way in several cities: Shenzhen,
Suzhou, Chengdu, Xiong'an and the “2022 Winter Olympics Office Area” in Beijing
(Cheng (2020)).
In China, the introduction of a CBDC should be seen in the context of a highly
digitised economy and widespread use of private digital payment services. The
introduction of a CBDC in the world’s most populous country and second-largest
economy may have far-reaching implications. In addition to providing a convenient
complement to cash for use in online transactions, a CBDC would also bring more
diversity to the current mobile payments duopoly of Alipay and WeChat Pay, which
collectively control 94% of the market for mobile payments (FSB (2019)). If there is a
decision to go beyond the current pilot stage, the DC/EP will become a complement
to M0, which includes banknotes and coins, as well as central bank depository
accounts. It is not intended to fully replace physical cash.
Graph 8 describes the main design characteristics of DC/EP, following the CBDC
pyramid. The architecture of the current DC/EP pilot is squarely the “hybrid CBDC”
model: it features a CBDC that is a direct claim on the PBC, but onboarding and real-
time payment services are operated by intermediaries (called “authorised operators”).
The central bank periodically receives and stores a copy of retail holdings and
transactions.
Fan (2020) emphasises that the role of the PBC is to provide the core
infrastructure, while intermediaries such as commercial banks, other payment service
providers and telecoms would provide services to the public. This approach prevents
concentration of risks at the central bank, disintermediation of existing financial
institutions, and duplication or waste of resources given the existing IT infrastructure,
processing capabilities and qualified staff at intermediaries.
The backbone of the DC/EP’s infrastructure would be a mixed system with
conventional database and DLT. PBC has, however, emphasised that DLT is not yet
sufficiently mature for such a large-scale application. To settle transactions, any
system has to be able to accommodate 300,000 transactions per second (TPS) to
accommodate the large retail transactions in China.

Design characteristics of the PBC’s DC/EP project (pilot phase) Graph 8
Sources: Adapted from Auer and Böhme (2020) and conversations with PBC staff.
The PBC does not require intermediaries to use any specific infrastructure or any
specific technological path. However, for transaction records and reconciliation, DLT
may be used. The financial intermediaries would remain responsible for know-your-
customer (KYC) checks and retail services. Nonetheless, the CBDC would be a direct
claim on the central bank, denominated in renminbi.
Regarding access, PBC has decided to use a value-based, semi-account-based
and account-based hybrid payment instrument. Identity would be based on “loosely
coupled account links”, such that users could use DC/EP anonymously with
counterparts in daily transactions, but that “operating agencies should submit
transaction data to the central bank via asynchronous transmission on a timely basis”
(Fan (2020)). This would ensure that users remain anonymous to each other, but allow
the central bank “to keep track of necessary data to implement prudent regulation and
crack down on money laundering and other criminal offences, as well as easing the
workload for commercial banks” (Fan (2020)).
Wallets are based on multiple forms of identification (ID), not all of which need
to include the name and other personal information. In particular, they could
accommodate tokens or accounts by intermediaries and allow individuals to decide
whether to connect to a bank account. To accommodate different levels of user
anonymity and access, there would be several grades of digital wallets based on the
strength of the KYC levels, with stronger KYC requirements associated with higher
transaction limits. Limits would generally be linked to existing rules of use of
banknotes and coins; details have, however, not yet been established.
Finally, regarding international interlinkages, the DC/EP would be connected
to existing retail and wholesale systems, including the RTGS system. The primary aim
for DC/EP is domestic retail use. Nonetheless, if an understanding can be reached
with foreign jurisdictions, non-residents (eg tourists and business travellers) could

access the DC/EP with a foreign cell phone number for an entry-level wallet. PBC
would work to ensure ongoing compliance with relevant AML/CFT rules.
Looking further into the future, the DC/EP could potentially be used for renminbi-
invoiced trade with foreign parties, but this is still subject to consultation with other
central banks and entities. With a proof-of-concept finished and the pilot under way,
the groundwork has been laid: initial testing involves the commercial banks, payment
service providers and other private sector institutions. All the authorised operators
have formulated an exit plan as part of the pilot programme, similar to a “sandbox”
model, to ensure that the process is reversible. The exact launch date remains
unclear. 21 In international discussions, representatives of PBC noted that cross-
country coordination can be useful to ensure consistent standards across borders.
Forums such as the CPMI and Financial Stability Board (FSB) were mentioned as
suitable settings for cooperating with other central banks, regulators and public
authorities.
Riksbank: the e-krona project
Another advanced CBDC project is that of Sveriges Riksbank, the world’s oldest
central bank. In Sweden, another highly digitised economy, cash use has been on the
decline for some years, to the extent that an increasing number of shops are no longer
accepting cash at all. Noting that its economy is witnessing “the greatest and fastest
decline in cash worldwide” (Riksbank (2019)), the Riksbank was at the global forefront
of discussing the possibility of issuing a CBDC (see Skingsley (2016); Ingves (2017)).
The Riksbank – like other central banks – has researched several technologies
and approaches (see Sveriges Riksbank (2017)). Currently, it is developing a proof-of-
concept of the e-krona project (see Sveriges Riksbank (2020)). Again, a CBDC would
be intended as a complement to, not a replacement for cash.
The architecture of the current Riksbank proof-of-concept is a Hybrid CBDC
(Graph 9). The CBDC is a direct claim on the Riksbank and payments are operated by
payment service operators. 22 On the specific design, Riksbank researchers Armelius
et al (2020b, p. 87) note that the ongoing pilot is a “decentralized database of all e-
kronor in circulation at any given moment, where the Riksbank verifies all transactions
before completion.” They classify this under the label “Decentralized solutions with
intermediaries”, noting also that such designs would require the Riksbank “to provide
a contingency solution if one or several intermediaries fail in order to prevent a
situation where a substantial numbers of end users are unable to make e-krona
payments” (see p 89).
The infrastructure and technical implementation are based on DLT using R3’s
Corda, to run with several notaries. Owing to the proof-of-concept nature, all are
currently under the control of the Riksbank and its technology partners.
When it comes to the access technology, the CBDC in the Riksbank’s pilot is
account-based, but an option for low-value token-based prepaid cards is considered.
In the ongoing pilot model, the Riksbank issues CBDC but they are stored in wallets
21
On 15 April 2020, screenshots of the testing wallet for Agricultural Bank of China (ABC) were
released in the media (Ledger Insights (2020)).
22
An earlier report on the e-krona had also considered the Direct option, labelled as “register based e-
krona with significant Riksbank commitment” (Riksbank (2017)).

at intermediaries. Access to the wallet is based on identifying the owner of the wallet.
Looking ahead, the Riksbank may also develop CBDC payment cards that can be used
directly for small value payments and without accessing a wallet (ie token-based
access).
It is also noteworthy that even with account-based access, the design will be such
that users remain anonymous vis-à-vis the Riksbank. On the one hand, intermediaries
are responsible for KYC and ongoing due diligence for each and every CBDC user. On
the other, the Riksbank only receives information on individual account balances and
payments, but no information on the actual account holders (ie the persons or firms
who are behind each account or payment).
Sveriges Riksbank’s e-krona (proof-of-concept) Graph 9
Sources: Adapted from Auer and Böhme (2020) and conversations with Sveriges Riksbank staff.
Beyond the current proof-of-concept, a wider range of designs are under

consideration. The Riksbank itself has not communicated a view on the preferred
technological architecture, other than that the e-krona would be a direct claim on the
Riksbank (ie the indirect/synthetic model is not considered). However, its researchers,
writing on their own behalf, have considered more detailed options. For instance,
when it comes to architectures, Armelius et al (2020b) discuss the policy trade-offs
from a centralised model without intermediaries, a centralised model with
intermediaries and a Synthetic (Indirect) model. On the latter, they argue that “it is
not clear if this should really be considered to be a CBDC”, as it is not a direct claim
on the central bank (Armelius et al (2020b), p 89). These researchers, however, also
judge that even for the Hybrid and Intermediated models, the level of involvement of
the Riksbank and the costs of running the system would still be substantial. Beyond
the architecture and the technological and access side, both conventional and DLT-
based technologies are being considered. Armelius et al (2020b) discuss the relative
merits of open source code and licensed proprietary technology. A key requirement

for the actual CBDC will be good offline properties – among others via the above-
discussed low value prepaid cards.
Regarding interlinkages, the focus on the e-krona is domestic, and retail use by
non-residents would only occur via the use of pre-paid cards by tourists for small
purchases. On the wholesale side, the CBDC will connect to the banking system and
also the RTSG, thus enabling cross-border payments.
Bank of Canada: CBDC as a contingency plan
The Bank of Canada has produced an outstanding body of research and policy
communication on the topic of digital currencies. Wilkins (2016) was among the first
high-level policy makers arguing for CBDC, and research by staff was actively
researching this and related issues early on. 23
Despite its early start, the Bank of Canada has not communicated that it is
developing a retail CBDC pilot or proof-of-concept. Instead, it has outlined a
comprehensive plan for the conditions under which Canada should develop a CBDC.
It also has set out potential architectures, and it has accumulated relevant technical
knowledge through a series of projects on novel payment technologies, also in
cooperation with other central banks (see, for example Bank of Canada (2017)).
Indeed, as noted by Lane (2020) in February 2020:
“we have concluded that there is not a compelling case to issue a CBDC at this
time. Canadians will continue to be well-served by the existing payment
ecosystem, provided it is moderni[s]ed and remains fit for purpose […] All this
being said, the world can change very quickly. The Bank of Canada can imagine
scenarios in which we would consider issuing a CBDC so we can continue to
provide Canadians with trustworthy methods of payment.”
In particular, the Bank of Canada has considered (i) a scenario in which the use
of physical cash is reduced or eliminated altogether, and (ii) a scenario in which a
private cryptocurrency or stablecoin makes substantial inroads as a means of
payment. In order to prepare for these eventualities, the Bank of Canada is conducting
stakeholder engagement discussions, working with universities and firms on the
design of a CBDC. It is also cooperating internationally as part of an international
working group on CBDCs.
If a CBDC is to be developed, the overall aim of the design is a digital claim on
the Bank of Canada that closely mimics the properties of physical cash, to the extent
this is possible. The CBDC would not replace cash, but designed as a digital addition
with advantageous resilience and accessibility features.
Because the overall design goal is sufficiently clear, the Bank of Canada has also
spelled out elements of the architecture (Graph 10). 24 For one, a CBDC would be “a
claim in Canadian dollars against the Bank of Canada” (Shah et al (2020)), ie the
Indirect/Synthetic approach is not being pursued. Rather, the analytical note lays out
23
See, among many others, Chiu and Wong (2014, 2015) on e-money, Fung and Halaburda (2016) on
a first assessment of the pros and cons of issuing CBDC and Kahn et al (2018) for an in-depth
investigation of desirability and associated design choices, and Davoodalhosseini (2018) for an
assessment of the implications for monetary policy.
24
Bank of Canada refers to the architecture as “Models” or “Business models” in all their
communications.

three potential architectures that correspond to the “Direct CBDC” (the Bank of
Canada providing the entire CBDC payment system) and the “Hybrid CBDC” (the Bank
of Canada only issuing and redeeming CBDC, with private sector intermediaries
providing end user services) or the “Intermediated CBDC” (identical to the Hybrid
model, where the Bank of Canada does not have access to the full ledger of retail
transactions). There is also the possibility of providing a Hybrid option in which
intermediaries execute the majority of payments, but the Bank of Canada can conduct
some retail payments directly in line with social goals. The goal of the latter approach
is to benefit from the valuable services offered by payment intermediaries, but at the
same time have available a directly operated option to serve public policy goals such
as universal access.
Bank of Canada’s CBDC contingency plan Graph 10
Bank of Canada
CBDC design
choices
Sources: Adapted from Auer and Böhme (2020) and conversations with Bank of Canada staff.
Details of the infrastructure have not been spelled out thus far. The Bank of
Canada has experience with a number of novel DLT-based proof-of-concept payment
projects. However, it notes that, whereas DLT may be possible as a solution for an
infrastructure, it is by no means necessary. Multiple technologies will be considered
simultaneously and winners will be picked based on performance.
Both account and token-based access solutions will be considered, likely with
tiering: anonymous token-based options (including but not limited to value cards)
would be allowable for smaller payments. This could be achieved through a low-cost
dedicated universal access device (UAD) (Miedema et al (2020)), which could allow
users without a smartphone to use the CBDC. Account-based access would be
required for larger purchases. As regards the details of such tiering, in particular when
it comes to enforcing proper AML/KYC standards, the current level of enforcement as
achieved with cash is the model. The tiering will be chosen such that AML and related
issues are not graver than is currently the case. Given the new options for illicit use

opened by programmable money, achieving the same level of enforcement might require a reduction in the
definition of what constitutes a “large” payment.25
On retail interlinkages, while the CBDC’s focus would be domestic, the option for token-based CBDC
naturally makes the system accessible to tourists and other non-residents for making low-value payments. Again,
the overall design of the system will be such that issues regarding international use and circumvention of
foreign capital controls are no graver than is the case in the current cash-based system. 26
On wholesale linkages, Bank of Canada will ensure that CBDC will be interoperable with all other
payment means, so that it can be exchanged freely with bank deposits and cash. The CBDC will also link directly to
the Large Value Transfer System, which handles wholesale payments. Due to the manifold interlinkages with the
domestic payment system, all linkages of these systems with foreign payment systems will also be guaranteed.
6. Conclusion
This paper has examined the rise of central bank digital currencies, a new payment technology that may soon be
available in a number of countries around the world. We have presented a novel CBDC project index (CBDCPI).
We have shown that this index is higher in jurisdictions with higher mobile phone usage and higher innovation
capacity. Especially retail CBDCs are more likely where there is a larger informal economy, and wholesale
CBDCs are more advanced in economies that have higher financial development. We have also noted that CBDC
projects differ starkly across countries, both in their motivations and their economic and technical design. Many
central banks are pursuing models where a CBDC is a direct claim on the central bank, but with private
intermediaries. To better understand these differences, we have zoomed in on three advanced cases, namely
those of the People’s Bank of China, Sweden’s Riksbank and the Bank of Canada.
Given the novelty of CBDC, and the scope for “clean-slate” thinking on the nature and provision of money, it is
natural that the approaches will differ across countries, in line with economic circumstances and users’ priorities.
In countries where digital payments are already very advanced, and cash use is declining, central banks may
respond in particular to ensure the ongoing availability of a public sector-provided means of payment. In
countries with a lower penetration of digital payments, financial inclusion may be an important driver. The choice
of architectures, infrastructures, access and interlinkages will be tailored to fit local circumstances.
Yet our overview has also shown some key common features. In particular, none of the designs we survey is
intended to replace cash; all are intended to complement it. Most still involve a strong role for intermediaries –
although potentially in parallel to direct provision of some services by central banks. None of the designs is pursuing
the Indirect model, where a CBDC is a claim on intermediaries rather than on central banks. We believe that by
sharing information on the drivers, approaches and technologies, central banks can learn from one
another, thus complementing international policy work in this area.
Going forward, events such as the Covid-19 pandemic highlight the value of access to diverse
means of payments, and the need for any payment method to be both inclusive and resilient against a broad range
of threats, just as cash is (see Auer et al (2020)). While it is difficult to anticipate the range of challenges ahead,
central banks will continue to take a long-term view and carefully consider the role of CBDCs in a range of potential
future scenarios.
25
The current definition of cash transactions reportable to FINTRAC is CAD 10,000 (in a single payment
or successive payments within 24 hours).
26
Currently, when travelling to Canada, amounts of foreign currency exceeding a value of CAN 10,000
or more must be reported to customs officials at the Canadian border.

Annex A: data collection for CBDC indicators
This annex gives a description of the construction of the CBDC project index (CBDCPI), and our central bank speech stance
and search interest scores.
The CBDCPI is constructed based on an update to the list of retail CBDC projects identified in Auer et al (2020a) with
the addition of a number of project announced after March 2020 and of wholesale CBDC projects. Each published
report can take a value of 1 for research reports, 2 if a pilot has taken place or is ongoing and 3 in the (so far
hypothetical) case of a live CBDC. We consider only jurisdictions that have a central bank or monetary authority. For
each jurisdiction, the overall index is the maximum of the retail and wholesale sub-indices. For China, given lack of
material translated into English and considering its prominent role in this field, we have defined a project score of
2. For Australia and the United States, we have defined a project score of 1, given the confirmation of such work in
reporting to the Australian Senate by the RBA (2019), and in Congressional testimony by Chairman Powell (2019) and in
speeches by Governor Brainard (2020a, b). For the euro area, given the work by both the ECB and several national
central banks of euro area countries, we include an observation for the euro area as a whole (with a project score of 1),
and each of the 19 euro area members (with 0 or 1 depending on whether the national central bank has published
any CBDC research). Empirical results are robust to dropping individual euro area members. For the Eastern
Caribbean Currency Union (ECCU), served by the Eastern Caribbean Central Bank (ECCB), the eight member states are
included as a single observation with a CBDCPI of 2 given the ongoing pilot. The countries of the West African
Economic and Monetary Union (WAEMU) are consolidated into one observation (CBDCPI of 0), as are the members
of the Economic Community of Central African States (ECCAS). Full links to public sources are available as part of the
background documentation.
The search interest score is estimated using a simple average of the interest score for the keywords “CBDC” (search
word) and “Central Bank Digital Currency” (topic) over the period January 2013–mid July 2020. The resulting two values
for each country are averaged to arrive at the sub-index. For China, we used the Baidu index for keywords “Central
Bank Digital Currency” and “DC/EP”. We rescaled the values to make them comparable to Google Trends figures (ie
values range between 0 and 100) and applied the same procedure described above.
The central bank speech stance score is obtained by classifying the stance of each central banker speech
containing at least one keyword from the following list: ”CBDC”, “Central Bank Digital Currency”, “digital currency”
or “digital money” (with a manual check to ensure it refers to CBDC and not private digital currencies). Speeches
come from the BIS central bankers’ speeches database (www.bis.org/cbspeeches/), a comprehensive
database collecting central bankers’ speeches as published on the BIS website for a wide selection of central banks
and international organisations. At the time of writing, the database counts 16,036 speeches, covers a period of
more than 23 years (1997-current) and has a wide geographical coverage (108 countries and 125 institutions). A
query yielded a set of 138 speeches that contained at least one of the keywords of interest. The resulting sample
covers the period December 2013–May 2020 and 38 countries including the euro area and several of its member
countries. When the speaker was an ECB official we labelled the speech as euro area. Conversely, if the speaker was an
official of a national central bank member of the Eurosystem we labelled the speech as the corresponding
country. For China, given the scarcity of material translated into English, we complemented the results with public
sources, including Fan (2020).
After compiling relevant speeches, we went through each and classified them by interpreting the stance of the
speech towards adoption of CBDC or CBDC more in general. Each speech score can take a value of either –1, 0 or +1
according to the specific speech stance. The score takes a value of –1 if the speech stance was clearly negative or in case
it was explicitly said that there was no specific plan at present to issue digital currencies. It takes a value of 0 in case of a
neutral stance. Finally, it takes a value of +1 if the speech stance was clearly positive or a project/pilot was launched or was
in the pipeline. Finally, the speech score was calculated as a simple average of the country level scores. If a country did not
have any speech score, we replaced the missing values with a zero, is in line with the interpretation of a neutral stance.
In the total sample of 175 countries or currency areas (ie euro area for ECB-related activities, Eastern Caribbean for the
Eastern Caribbean Central Bank and WAEMU and ECCAS for these currency unions), 55 had a non-zero value for either
the CBDCPI or one of the two underlying scores. In the other 120 countries or currency areas without any communication
on CBDC, CBDCPI takes the value of 0.

Annex B: overview of CBDC projects
Selected retail CBDC projects Table B1
Design choices
Infrastructure2
Interlinkages4
Architecture1
Project/country Notes on status, motivation and conclusion

Access3
Research; Aim to address “steadily diminishing use of banknotes and

Rafkróna
D U A N coin”; “many issues have yet to be clarified, and they must be dealt
Iceland
with appropriately before a position can be taken”.
Pilot; Improve “financial inclusion […], [reduce] the size of legitimate
but unrecorded economic activities, [strengthen] national defences
Sand Dollar
D U A N against money laundering and other illicit ends [and][…] deliver
The Bahamas
government services through digital channels, thereby improving tax
administration and increasing the efficiency of spending”.
Research; “the potential benefits of introducing CBDC are not
“E-krone*”
D U U N assessed to match the considerable challenges that the introduction
Denmark
would present”.
Working group; focus on “independent back-up solution, credit risk-
“E-krone*”
D U U N free alternative to bank deposits, competition, legal tender”; “more
Norway
information is required before a conclusion can be reached”.
Research; “The introduction of CBDC does not require the use of
DLT.”; “opting for CBDC implies trust in the central bank, which is why
“E-euro*” we have chosen a reference design with a centralised consensus
H or I C A I
The Netherlands mechanism.”; “the technology behind account-based digital payment
systems has proven itself to a far greater extent than value-based
digital money”.
DC/EP (Digital
Currency/Electronic Ongoing work; aim to create digital alternative to cash and coins for
H or I D & C A N
Payments) retail use.
China
Research; “CBDC with the status of legal tender could guarantee that
“E-euro*” all users have, in principle, access to a cheap and easy means of
H or I D A/T I
ECB payment“; “proof of concept also highlights a number of areas where
there is room for improvement”.
Contingency plan; “building, as a contingency, the capability to issue a
“E-dollar*”
H or I D & C A/T N cash-like central bank digital currency (CBDC) to the public, should
Canada
the need ever arise.”
Pilot; “within a few years, if the current trend continues, we will find
ourselves in a situation where cash is no longer generally accepted as
E-krona
H or I D A/T N a means of payment”; “an account-based e-krona could rationalise
Sweden
payments from agencies and make them less dependent on
commercial agents”.
Research; “Improve the efficiency of the monetary function, …
payment processes and systems, …. financial inclusion and … user
Digital Fiat Currency
H or I D T N experience”.
Brazil
Architecture: Hybrid, as on the one side “any relationship between the
Central Bank and society is intermediated by ﬁnancial institutions, just

as with physical cash.” On the other side intermediaries are only
“custodian of the new cash form”.
Research; “households and businesses should be able to make fast,
efficient and reliable payments, and benefit from a resilient, inclusive,
innovative, and competitive payment system.”; “two key elements of
“E-pound*”
H or I U U U the platform: (1) a core ledger, provided by the Bank, would record
United Kingdom
CBDC and process payments, and (2) private sector ‘Payment Interface
Providers’ would handle the interaction with end‑users of CBDC and
provide additional payments functionality through overlay services”.
Pilot; Aim to address the “high cost of current payment instruments
DXCD
U D T I and banking services”, needs of customers and inefficient cheque
Eastern Caribbean
settlement.
Dinero Electrónico Pilot; “means of payment available to absolutely all Ecuadorians”.
U C A N
Ecuador Operated 2014–16; discontinued.
Bakong Pilot; aim to “increase access to quality formal financial services”;
U D U N
Cambodia “decrease demand for… cash”.
Pilot; Test DLT “as a technological framework for e-hryvnia issuance
E-hryvnia
U D U N and circulation”; no fundamental advantage in using DLT in a
Ukraine
centralised model.
Electronic legal Expression of interest; “The scope of this project is specific to the use
U U T N tender of a CBDC as electronic legal tender (ELT), similar to the characteristics
South Africa of, and complementary to, cash.”
Research; “[…] account based model would offer better results for a
“E-euro*”
U U U I retail CBDC. However, it might also lead to a greater loss of resources
France
for banks […]”.
Research; “The case of non-anonymous CBDC based on technology
“E-euro*” similar to the current electronic payment methods would imply
U U U I
Spain significant infrastructure costs and operational and regulatory
requirements”.
Billete Digital Pilot; “Digital bills that aim to have same functions and uses as
U U U N
Uruguay physical bills”; ongoing evaluation.
Research; “help in the struggle against unreported transactions”;

E-shekel “contribute to the high-tech sector (fintech)”; Conclusion that “the
U U U N
Israel team does not recommend that the Bank of Israel issue digital
currency (e-shekel) in the near future”.
Research; “Safer and cheaper to transport than cash”; “Provides public
access to an electronic form of legal tender”; “Reduces cash demand
CBDC series
U U U N and supply which could reduce the availability of cash in an electricity
New Zealand
outage”; “Slow payment authorisation in a blockchain-like crypto-
currency”.
Research; “evaluate whether clear use cases exist for the introduction
of a retail and/or wholesale CBDC specifically within the context of
“E-lilangeni*”
U U U N Eswatini. Findings from the first phase of investigation indicate that
Eswatini
there are indeed at least three potential use cases”. “However, while
these results are positive, further research is warranted…”.
“E-rupiah*”
U U U N Research.
Indonesia
Research; “Pioneering work reviewed generally concludes that CBDC,

“E-ringgit*” even if introduced in the future, would likely be a complement rather
U U U N
Malaysia than a substitute to cash and bank deposits”. “Technological hurdles
need to be considered, as central bank credibility must take priority”.

“E-won*” Pilot; Start researching and reviewing technology, process, legal
U U U U
Korea framework. Test if CBDC will technically work in a limited environment.
“E-dinar*”
U U U U Research.
Tunisia
“E-rouble*”
U U U U Research.
Russia
Research. “…the opportunities and challenges of, as well as the use

“Digital-dollar*”
U U U U cases for, a CBDC, as a complement to cash and other payments
United States
options”.
Research. “… the case for issuing a CBDC for use by households has
“E-AUD*”
U U U U not been established”; “… the implications of CBDC for the structure
Australia
of the financial system would need to be carefully considered”.
“Digital-Yen*”
U U U U Research.
Japan
Research; “Examine the opportunities and risks of introducing a

E-franc
U U U U cryptofranc (e‑franc)”; “additional benefits currently low, outweighed
Switzerland
by risks”.
1
D = Direct; H or I = Hybrid or Intermediated; U = unspecified or multiple options under consideration. 2 C = Conventional; D = DLT; U =
unspecified or multiple options under consideration. 3 A = Account-based; A/T = tiering of Account- and Token-based; T = token-based;
U = unspecified or multiple options under consideration. 4 I = International; N = National; U = unspecified or multiple options under
consideration. * Not an official designation by the central bank.
Sources: central banks’ websites; conversations with central bank staff; www.centralbanking.com; www.unescap.org; www.efd.admin.ch.

BIS Working Papers: No 880 Rise of The Central Bank Digital Currencies: Drivers, Approaches and Technologies

Uploaded by

Copyright:

Available Formats

BIS Working Papers: No 880 Rise of The Central Bank Digital Currencies: Drivers, Approaches and Technologies

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

BIS Working Papers: No 880 Rise of The Central Bank Digital Currencies: Drivers, Approaches and Technologies

Uploaded by

Copyright:

Available Formats

BIS Working Papers

Monetary and Economic Department

JEL classification: E58, G21.

Keywords: central bank digital currency, CBDC, central

Raphael Auer,♦ Giulio Cornelli♦ and Jon Frost♦†

Rise of the CBDCs: drivers, approaches and technologies 1

Rise of the CBDCs: drivers, approaches and technologies 3

Rise of the CBDCs: drivers, approaches and technologies 4

2. Taking stock of CBDC research and development efforts

Rise of the CBDCs: drivers, approaches and technologies 5

Rise of the CBDCs: drivers, approaches and technologies 6

Sources: Central bankers’ speeches; central banks’ websites.

As of mid-July 2020, at least 36 central banks have published retail or wholesale

CBDC projects status Graph 3

Source: central banks’ websites.

Rise of the CBDCs: drivers, approaches and technologies 7

Rise of the CBDCs: drivers, approaches and technologies 8

General purpose (retail) CBDC Wholesale CBDC

1 = not so important; 2 = somewhat important; 3 = important; and 4 = very important.

Sources: CPMI survey of central banks; Boar et al (2020).

3. The cross-country drivers of CBDC development

Several global developments – including the digitalisation of commerce, the rise of

Rise of the CBDCs: drivers, approaches and technologies 9

A novel CBDC database

• A search interest index corresponding to Google Trends or (for China) Baidu

For each of the indicators described above, we replace country-level missing

Rise of the CBDCs: drivers, approaches and technologies 10

Pairwise correlations Table 1

CBDC project index1 Central bank Search interest index

Examining the cross-country drivers of CBDC projects

Rise of the CBDCs: drivers, approaches and technologies 11

Rise of the CBDCs: drivers, approaches and technologies 12

Descriptive statistics1 Table 2

Rise of the CBDCs: drivers, approaches and technologies 13

Rise of the CBDCs: drivers, approaches and technologies 14

Rise of the CBDCs: drivers, approaches and technologies 15

Wholesale CBDC project

Retail CBDCs, also appear to be more advanced in jurisdictions with high

4. Policy approaches and technical design

Rise of the CBDCs: drivers, approaches and technologies 16

A stocktaking framework: the CBDC Pyramid

Approaches to CBDC design are heterogeneous across countries, requiring us to distil

The CBDC pyramid Graph 6

Source: Auer and Böhme (2020).

Rise of the CBDCs: drivers, approaches and technologies 17

Rise of the CBDCs: drivers, approaches and technologies 18

Attributes of retail CBDC projects

Architecture Infrastructure Access Interlinkages

Sources: central banks’ websites; authors’ calculations.

Regarding infrastructure, we find seven central banks running their prototypes

Rise of the CBDCs: drivers, approaches and technologies 19

The drivers of technological designs

Rise of the CBDCs: drivers, approaches and technologies 20

Architecture Infrastructure Access

Government effectiveness 0.599**

5. Approaches for CBDC design: three examples

Rise of the CBDCs: drivers, approaches and technologies 21

PBC: the DC/EP project (pilot stage)

Rise of the CBDCs: drivers, approaches and technologies 22