Digital Money & Central Banks Balance Sheet

Digital Money & Central Banks Balance Sheet

The creation and evolution of central banks is a landmark in the long history of financial markets
and payment systems. The universal use of mobile phones and the platform revolution has
created the conditions for a financial revolution. Digital money is the next step in the process of
continuous technological advancement in payment systems. In response, central banks are
reviewing their conduct of monetary policy and operations in light of the new topology of
financial markets and payment systems. This paper studies the potential implication of digital
money on central banks balance sheet.
For a long time, society chose a metal (i.e., gold) as money, and governments facilitated the
payment mechanism by producing coins with a standardized weight. Banks further contributed to
the process by opening accounts (private money) at one-to-one exchange rate with gold (or
silver). These accounts reduced both the risk of theft and transaction costs, in particular for large
amounts. However, there was no central bank in charge of regulating the money supply, since
using a metal as currency was the norm; banks only provided payment services under a narrow
banking structure. The latter provided security against liquidity shortages, as deposits were
backed 100 percent by gold (or silver). The aggregate money supply was the amount of available
gold (or silver).
A significant change took place when banks engaged in lending activities by the time of the
Industrial Revolution, as they realized that deposit withdrawals were not simultaneous, and
therefore it was not necessary to keep a gold/silver backing of 100 percent; (i.e., a reserve
requirement of 100 percent). This gave birth to bank credit and maturity transformation of
financial assets. Bank lending (was envisaged) to increase potential GDP by allowing more
efficient allocation of savings to economic activities through financial intermediaries. The
Federal Reserve System was created in 1913, as a result of crisis and thus there was a need for
lender of last resort (LOLR), with a mandate to provide elasticity to the currency.
The collapse of the Bretton Woods exchange rate peg system in 1971 and the high inflation
episode during the 1970s and early 1980s—with central banks in the developed economies
operating under a floating exchange rate system—paved the way for a new monetary policy
framework. However, together with these macroeconomic developments, there were also
important improvements in payment systems, which induced changes in monetary practices.
Financial innovations, such as debit cards and automated teller machines (ATMs) during the
1970s and 1980s, facilitated payments from bank accounts; demand for money became unstable.
The traditional transaction demand for money—based on the notion of “shoe-leather costs”
(Baumol, 1952) became irrelevant, as ATMs could be found everywhere, and credit cards
INTERNATIONAL MONETARY FUND

5

IMF WORKING PAPERS

Digital Money and Central Banks Balance Sheets

replaced currency in circulation (CiC). To keep monetary aggregates as an intermediate target in
the context of higher substitution between CiC and savings deposits, central banks began to use a
broader definition of money. In the end, unstable demand for money was not conducive for
monetary targeting in low-inflation countries and their central banks adopted interest rate as
operational target —including those adopting inflation targeting (IT). As the former central bank
of Canada’s Governor Gerald Bouey put it (1983): “Central banks did not abandon monetary
targeting; it was the other way around: money demand left central banks.”
As of today, the impact on digital currencies differs across countries depending on the status of
their respective currencies. The challenge for central banks is to preserve the efficiency of
monetary policy and adapt to the impending wave of digital money.
In section II we describe the main features of bi-monetary (or partially dollarized economies),
with a focus on the functions of money and the reason for the hysteresis phenomenon. In section
III, we discuss how central banks balance sheet and monetary policy may change under the new
environment of digital money, including the issuance of CBDC. Section IV discusses the
implementation of digital money and new costs that may accrue to the central bank from cyberrelated risks, (distinguishing between wholesale and retail costs). Section V shows six
simulations of how banking loans, money and central banks balance sheet may change for
different types of substitution with digital money (e.g., stablecoins, MNOs, and CBDC), and
complements recent IMF policy work in this area (e.g., IMF (2021); IMF (2020)). Section VI
concludes with some forward-looking suggestions.

II. Lessons from Bi-Monetary Systems
Most economies operate with a foreign exchange (FX) (e.g., the dollar) bias for international
trade and finance invoicing. Additionally, banking systems in many developing economies are
bi-monetary. This section reviews the main features of partially dollarized economies, with a
focus on the functions of money as unit of account, means of payment, and store of value.
Table 1 describes four types of countries according to the functions of local currencies. On one
end, the U.S. enjoys a privileged status as issuer of the most widely used international currency
(Type A); and on the other end, countries like Venezuela and Zimbabwe have very limited
monetary room to maneuver due to high dollarization (Type D). It is a simple characterization of
current status, and any country could combine some features of different type-cases of Table 1. It
may be useful to use dollarization literature in assessing crypto assets. As the latter is not money,
but performs some of its functions, there is imperfect substitution among them.

INTERNATIONAL MONETARY FUND

6

IMF WORKING PAPERS

Digital Money and Central Banks Balance Sheets

Table 1. Countries Classified by Dollarization Level
Country/Function of
Money

Unit of Account

Means of Payment

Store of Value

Issuer of International
Currency (Type A)

Fiat currency used in
domestic and most
international trade
transactions.

Fiat currency used in
domestic and most
international trade
transactions.

Fiat currency used in
domestic and international
lending/securities
operations.

Open Economy under
Dominant Currency
Paradigm (Type B)

Fiat currency used in
domestic transactions. FX
used in foreign trade.

Fiat currency used in
domestic transactions. FX
used in foreign trade.

Fiat currency used in
domestic financial contracts.
FX used in international
lending/securities operations
(except for sovereign bonds
in certain cases).

Partially Dollarized
Economy (Type C)

Fiat currency used in
transactions of final goods
and services. Partial
dollarization of prices of
durable goods and
tradable inputs. FX used
in foreign trade.

Bi-monetary system, with
pricing denominated in
the fiat currency. Tax
payments in the fiat
currency. FX used in
foreign trade.

Bi-monetary system, where
FX is used to diversify risks
in turbulent episodes. FX
used in international
lending/securities operations
(except for sovereign bonds
in certain cases).

Highly Dollarized
Economy (Type D)

Extended use of the
exchange rate for price
indexation (high real
dollarization and almost
complete pass-through
from depreciation to
inflation). FX used in
foreign trade.

Limited scope for fiat
currency (tax payments,
public expenditure, nondurable goods, and lowvalue transactions).
Extended FX use for
durable goods, real estate,
capital goods, and highvalue transactions.

FX takes over the role of
store of value. Lending
capacity in domestic
currency becomes limited.
Most loans become FXdenominated when FX bank
deposits are allowed.

A bi-monetary system embodies the failure to conduct monetary policy in an effective way,
i.e., secure price stability, efficient payment systems, and well-functioning financial markets
(including long-run financial contracts at comparatively low nominal interest rates). Particularly,
under high and persistent inflation, market participants defend themselves by shifting to FX.
The most common type of dollarization is financial dollarization (FD), or asset substitution,
caused by a poor performance of the local currency. The local currency is used more for payment
INTERNATIONAL MONETARY FUND

7

IMF WORKING PAPERS

Digital Money and Central Banks Balance Sheets

transactions but is replaced by the dollar as saving asset or store of value, in line with Gresham’s
law. Under extremely high inflation, such as in Venezuela or Zimbabwe, real dollarization
(RD)—i.e., use of the dollar as means of payment transactions and store of value—also takes
place (Table 1, Type D).
On the one hand, in some countries dollarization is entrenched and a bi-monetary system is
formally allowed as it is described in Table 1 for Type C (e.g., Uruguay). On the other hand, in
other countries it is not allowed, or dollar accounts are restricted. Under high inflation (e.g.,
Argentina or the Democratic Republic of the Congo), the public holds a large share of financial
assets abroad and local financial intermediation is low. Countries with no history of extreme high
inflation (e.g., Malaysia) impose restrictions on dollar deposits, but there seems to be no
significant impact on local financial intermediation (Table 1, Type B).
Once a country gets used to a bi-monetary system, the process is not easy to reverse, even when
the initial trigger (e.g., high inflation) subsides, a phenomenon known in the literature as
hysteresis. The optimal choice between domestic currency vs. FX will depend on the monetary
framework and the benefits that each may offer as they co-exist as two currencies.
From the recent performance of bitcoins (and the like), and stablecoins vis-à-vis the functions of
money, it is apparent that stablecoins has significant implications for monetary policy design and
liquidity management, as a fast-growing stablecoin industry may induce financial
disintermediation (Kahn and Singh, 2021). Furthermore, stablecoins backed by central bank
reserves allow money supply control with the central bank; this control with the central bank will
be diluted, if backed by anything other than central bankreserves. In the case of international
currency issuers (U.S. and the Eurozone), higher demand for bonds (e.g., U.S. Treasuries or
German Bunds or AAA private bonds) to back stablecoins might result in larger capital markets
and smaller banking systems. 1 In other countries, financial disintermediation might take place
through capital outflows, as the expansion of stablecoins linked to an international currency may
increase dollarization or euroization. 2
A bi-monetary system limits the role of the exchange rate as a shock absorber, as RD implies a
high pass-through from exchange rate depreciation to inflation. FD creates currency mismatches

Kahn and Singh (2021) argue that reserves at the central bank may be the best option to keep stablecoins “stable”;
this is preferred to (for example) U.S. Treasuries as collateral as they will be siloed and adversely impact market
plumbing; central bank reserves are also more liquid than U.S. Treasuries for T0 settlement. See: If stablecoins are
money, they should be backed by reserves – Risk.net.
2
A similar outcome is expected if a retail CBDC of an international currency is developed, but as of today it is
unclear that the Fed or the European Central Bank (ECB) will choose to do that in the near future.
1

INTERNATIONAL MONETARY FUND

8

IMF WORKING PAPERS

Digital Money and Central Banks Balance Sheets

and liquidity risks for the financial system and the economy as a whole. Therefore, the exchange
rate amplifies negative external shocks rather than absorbing them. Both FD and RD jeopardize
monetary transmission mechanisms, as inflation expectations are difficult to anchor with a weak
interest rate channel. FD-related financial instability would need to be addressed via policy
responses such as a central bank FX reserve buildup and associated regulation.
The risk of higher dollarization induced by dollar-denominated stablecoins will be larger in
countries with high inflation, capital control measures, and/or restrictions to open accounts in
foreign currency within the local banking system.

III. Monetary Policy and Central Banks Balance
Sheet
Monetary policy actions usually go beyond monetary procedures to reach the operational target
or discount window operations as a LOLR. In recent years, central banks in developed
economies started to expand their balance sheet in the aftermath of the Global Financial Crisis
(GFC), and central banks of developing economies did the same, before resorting to international
reserves and other balance sheet operations. 3
Under a monetary dominance regimen, a central bank has full control on its balance sheet under
its legal mandate of price stability (or dual mandate in some countries). The profit or losses made
by the central bank pursuing its own mandate is taken as a given for the fiscal authority who is
responsible for keeping public sector finance sustainable overtime. The equation from the
classical paper of Sargent (1981) about the unpleasant monetaristic arithmetic makes this point:
∞

∞

=
bt −1 R −1 ∑ R − i stf+i + R −1 ∑ R − i st +i
=i 0=i 0

The equation essentially says that the net present value (NPV) of the primary fiscal surplus plus
the NPV of seigniorage must be equal to public debt. Therefore, any changes of seigniorage
under monetary dominance must be accommodated by a change in the fiscal position. On the
contrary, if the digital money revolution occurs under a fiscal dominance regimen and reduces
seigniorage, central banks may be inclined to take actions to avoid that situation by raising
According to Ceccheti (2008), the management of a central bank’s balance sheet has two general principles. The
first is that a central bank controls the size of its balance sheet, and by doing so can affect the level of the risk-free
interest rates. The second is that a central bank controls the composition of its assets. Depending on that
composition, central banks may influence relative prices of bonds and exchange rates.

3

INTERNATIONAL MONETARY FUND

9

IMF WORKING PAPERS

Digital Money and Central Banks Balance Sheets

inflation tax or preventing the development of financial innovations. From a welfare perspective,
it is preferable to have a regime of monetary dominance or central bank autonomy as the price
stability goal than a fiscal dominance regime where central bank is not able to achieve its
mandate because of lack of operational independence (Leeper, 1991).
Payment Systems and Seigniorage: If seigniorage is not a constraint (i.e., profits to be transferred
to the Treasury), the development of a payment system should focus on how to improve its
efficiency. Regulation should encourage a competitive environment of competition where tech
firms can innovate, and citizens benefit from these new technologies. For all participants
(newcomers and incumbent ones), a level playing field, or the principle of “same regulation for
the same type of financial activity or risk”, should apply to avoid regulatory arbitrage. Regulators
should embrace positive externalities wherever possible (e.g., interoperability networks). Only in
cases where private sector does not find it profitable to invest in externalities, should the state
step in.
From a monetary policy perspective, the interest rate and credit channel may become weaker in
the cases where: (i) a Fintech issues stablecoins backed by treasury bonds; and (ii) substitution of
bank deposits by stablecoins in foreign currency. The common factor in these cases is the
reduction of banking credit. On the contrary, when CiC is substituted for private digital money
that is deposited in the commercial banks, there is an expansion of the banking credit—see
Section V for six illustrative cases of how digital money will impact central banks balance sheet
line items.

IV. Base Money Issues: M0 or Not
Any CBDC would coexist with other forms of money (so not an exclusively M0 world). The
question is what the liability structure even if being distributed by banks (or nonbanks). One state
of the world is where someone would have a traditional or bank tokenized deposits (M1), and
some balances as CBDC (M0). Or, the bank holds all the M0 (like they do with reserves today)
and use that to offer tokenized deposits (M1). The value added of a CBDC may lie in the ability
to establish a token platform that enables new technologies to do things that cannot be easily
done in traditional systems. The distribution model and account of M1 and M0 could look the
same (or different) but the platforms may offer alternatives (see also Adrian and Mancini Griffoli
(2019) on synthetic CBDC; Mancini Griffoli et al. (2017)).
Figure 1 shows a standard balance sheet with M0 depicted in a green shade (typically, CiC and
required reserves; lately some advanced economies due to QE have excess reserves also). The
importance of CiC is illustrated by Bindseil, in his 2016 Jackson Hole speech:
INTERNATIONAL MONETARY FUND

10

IMF WORKING PAPERS

Digital Money and Central Banks Balance Sheets

“[An] outstandingly lean central bank balance sheet was the one of the Fed pre-crisis, where the
total balance sheet length was only around 1.1 times the total amount of bank notes in
circulation. [ ]…the idea that the central bank permanently injects monetary accommodation
through a longer balance sheet with substantial holdings of a portfolio of less liquid assets with
long maturity and possibly some credit riskiness does not appear sufficiently convincing.”
Figure 1. Standard CB Balance Sheet (Focus on CiC and Monetary Base)

Source: IMF staff

Figure 2 shows central bank issued CBDC or a bank/non-bank issued stablecoins counted as part
of M0—think of access to central bank master account or payment rails. The central bank
acknowledges liability ex- ante. So “10” is part of the central bank liability. However, the right
side of the figure shows bank/non-bank issued stablecoins, but not a CBDC; so, not a central
bank liability ex-ante. In this scenario, stablecoins worth “10” will be outside M0; this will be in
the liabilities side of banks/nonbanks, as banks/nonbanks issued them by debiting their excess
reserves at central bank (preferably ringfenced to back-up their coins). Successful cases of
payment interconnection, such as Pix in Brazil and UPI in India, are examples of that trend.
Figure 2. Issuance of CBDC or Stablecoins by the CB (left) or Outside the CB (right)

Source: IMF staff

INTERNATIONAL MONETARY FUND

11

IMF WORKING PAPERS

Digital Money and Central Banks Balance Sheets

More importantly, if the basic features of CiC hold (non-interest bearing, no caps, and complete
privacy), then CBDC is digital CiC. CBDCs may also play a role in the process of improving
payment systems and supporting financial inclusion in the near future. A recent speech by the
Reserve Bank of India (RBI) articulates the CiC and CBDC relation very well: 4
“To sum up, CBDC is the same as currency issued by a central bank but takes a different form
than paper (or polymer). It is sovereign currency in an electronic form and it would appear as
liability (currency in circulation) on a central bank’s balance sheet. The underlying technology,
form and use of a CBDC can be moulded for specific requirements. CBDCs should be
exchangeable at par with cash.”
However, if CBDC deviates from CiC (e.g., interest bearing and/or capped), then it is a different
instrument (a policy tool) that arguably targets a new objective. 5 This will be a subtle issue
(e.g., as an illustration, if Eurozone’s caps CBDC for retail at €1 trillion, with the underlying
assumption that caps will not be binding as average retail transactions are around €1 trillion);
also see Annex I. 6 Also, digital money velocity (instantaneous payment) will be faster than CiC
(Kahn et al., 2002). Thus, a mix of CiC and CBDC in time t+1, may be less than CiC in time t=0;
base money may decline; and then so will seigniorage.
For emerging markets, Brazil is at the frontier. The basic idea of Brazil’s potential for adopting a
CBDC are laid out in Araujo (2022). The paper focusses on CDBD under the premise that it
should, in order to be useful, go beyond what an instant payments system should do. 7 The
underlying economic argument by Brunnermeier and Payne (2022) that describes an economy
where a smart CBDC arises as a new instrument for regulation in a digital assets environment

4

Reserve Bank of India – Speeches (rbi.org.in), July 2021 by Deputy Governor T. Rabi Sankar.

Armelius et al. (2018), suggest non-interest-bearing CBDC supplied according to demand would diminish the
impact on a negative policy rate as economic agents now have the chance to demand CBDC which is a closer
substitute of financial assets (for example, demand deposits) than CiC.
6
Visa, Mastercard, and related credit card use has been just over 1 trillion for the eurozone. Visa and Mastercard
provide transaction volumes: for example, Q3FY22-Visa-Operational-Performance-Data-FINAL-v2.pdf
(q4cdn.com); Page 1, Visa USD 661b; similar numbers for Mastercard (around 600 bn).
7
The primary use for a digital Brazilian Real is to serve as the basis for a smart settlement platform. The proposed
basic architecture, tokenized deposits (in banks or other PSP) constitute a form of private money, effectively
fulfilling the role currently played by stablecoins. Digital services could, thus, be built on that form of money.
Constraints on unannounced conversion flows (e.g., USD/period of time) among CBDC and other private monies,
could suffice to avoid bank runs.
5

INTERNATIONAL MONETARY FUND

12

IMF WORKING PAPERS

Digital Money and Central Banks Balance Sheets

(new objective)—is in line with the Tinbergen rule. 8 If the underlying economics argue for a
system to build financial services on top of the digital real instead of using their own
“stablecoins” then there will be need to reflect the change in the central bank balance sheet.
Thus, at present, Brazil’s Pix system, an instant payments system, is not part of the M0 world.
However, if Brazil adopts CBDC, (assuming benefits are significantly more than Pix offers) then
the central bank embraces the additional “10” liability (i.e., 100 in Figure 2, left panel; not 90 as
in right panel).
Costs from cyber-risk: Analytically, although the central bank liability is lower if issuance of
quasi-CBDC/stablecoins is outside the central bank, ex-ante, large costs “C” (i.e., systemically
important) from cyber-risk will be picked up by the regulator, due to financial stability and the
too-big-to-fail logic. 9 Thus, it is likely that in wholesale CBDC literature, C will be absorbed by
the central bank (or ministry of finance/fiscal authority), irrespective of the issuer. If the costs are
small “c” (e.g., as in retail CBDC issuance) then private vendors (i.e., PSPs), or banks/nonbanks/
Fintechs may pick up the cost, c.
This cost issue (C vs. c) is germane to those central banks contemplating both retail CBDC and
wholesale CBDC, especially interoperability between the two. The discussion is preliminary but
when the economics and technology of interoperability converges, C and c will not be separable.

V. Monetary Aggregates and Bank Credit Under
Digital Money
In this section, we discuss how banking credit and money may change under this new
environment of digital money. Our starting framework is a central bank with a policy rate as
operational target and we will see the impact of different development of digital money
depending on the type of money substitution that take place.

Tinbergen rule: Macroeconomic variables the policy maker wishes to influence are targets; instruments are
variables that the policy maker can control directly. Achieving the desired values of a certain number of targets
requires the policy maker to control an equal number of instruments.
9
Cyber risk and associated costs much higher than counterfeit paper currency that (in the literature) is not more than
0.01 percent of currency issued. (The costs referred to here have no reference to lower printing cost as fiat money
reduces in favor of CBDC.) In the US, Fed is required by law to do full cost recovery for all payments services. This
puts the Fed on an even playing field with private sector PSPs. In Euroland, the ECB does full cost recovery minus a
“public good” adjustment; thus priced below private sector.

8

INTERNATIONAL MONETARY FUND

13

IMF WORKING PAPERS

Digital Money and Central Banks Balance Sheets

The cases to study are substitution of:
(i) Bank deposits by stablecoins or e-money (assets outside the banking system) backed by
banking deposits. 10
(ii) Bank deposits by stablecoins or e-money (assets outside the banking system) backed by
treasury bonds or similar assets (i.e., good collateral).
(iii) Bank deposits by dollar-denominated stablecoins in a partially dollarized economy.
(iv) CiC by e-money.
(v) CiC by CBDC.
(vi) Bank deposits by CBDC (similar to reserves at central bank).
Case (i) Substitution of Banks Deposits by Stablecoins or E-money backed by Banking
Deposits
To illustrate this case, our starting point is period 0 when we have a traditional banking system
framework with a reserve requirement (RR) of 5 percent on deposits with zero excess of
reserves. In order to diversify credit and liquidity risks commercial banks hold 10 percent of
deposits at current interest rates at time t=0. 11 At the current interest rates, the preference for CiC
(CiC/Money) is 25 percent. M1, M0 and deposits are equal to 4000, 1150, and 3000
respectively. 12 Therefore, the central bank’s balance sheet at time t=0 is:
Commercial Bank (t=0)
Assets
2550 Commercial Loans

Liabilities
3000 Individuals’ Demand Deposits

300 Treasury Bonds
150 Reserves at Central Bank

For example, USDC.
The paper of Bernanke-Blinder (1988) determines a credit channel of monetary policy in a model where banks
choose the composition of portfolio among banking reserves, loans and treasury bonds depending on the relative
yield between loan rates and bonds and the RR conditions. However they do not consider net worth of the CB.
10
11

For illustration we take CiC/M1; any other denominator can also be used (e.g., M0 or the size of the balance
sheet).
12

INTERNATIONAL MONETARY FUND

14

IMF WORKING PAPERS

Digital Money and Central Banks Balance Sheets

Central Bank (t=0)
Assets

Liabilities

1150 Interest Bearing Financial Assets

1000 CiC
150 Reserves

In period t=1, a mobile operator phone company starts to provide e-money and by regulation (or
its own decision), the float that the company captures is deposited at a commercial bank. We
assume that this e-money improve payments made through the banks, so there is no substitution
with CiC at the central bank; only a change in banking deposits that breaks 3000 at t=0 into 2800
(individual demand deposits) and 200 (mobile phone company deposits).
Mobile Operator Phone Company (t=1)
Assets

Liabilities

100 Infrastructures

100 Capital

200 Deposits with Commercial Banks

200 Customer Accounts

Now the balance sheet of the commercial bank looks like this in period t=1:
Commercial Bank (t=1)
Assets

Liabilities

2550 Commercial Loans

2800 Individuals’ Demand Deposits

300 Treasury Bonds

200 Phone Company Demand Deposits

150 Reserves at Central Bank

In this case, there is only a redistribution of demand deposits and no changes of the asset side.
The central banks balance sheet does not change at all. 13

As core elements and regulators of traditional payment systems, central banks have also encouraged more
integrated systems, mostly via interoperability, particularly in developing economies, e.g., Pix in Brazil and UPI in
India. Pix, a platform created by the Central Bank of Brazil (BCB), interconnects multiple payment systems.
(continued…)
13

INTERNATIONAL MONETARY FUND

15

IMF WORKING PAPERS

Digital Money and Central Banks Balance Sheets

Case (ii) Substitution of Banks Deposits by Stablecoins or E-money Backed by Treasury
Bonds (i.e., good collateral)
Let us now see the case of a stablecoin with a policy of keeping 100 percent of treasury bonds as
backup of its liabilities (period t= 1a). Our starting point is again period t=0.
Fintech Issuer of a Stablecoin (t=1a)
Assets

Liabilities

100 Infrastructures

100 Capital

200 Treasury Bond

200 Stablecoins

This a case of financial disintermediation. Fintech is increasing the demand for treasury bonds
and commercial banks face a withdrawal of 200. Demand for bank reserves falls by 10 (i.e.,
5 percent of 200); other market players reaction matter. In the very short term, we can assume
that commercial banks will sell bonds to the Fintech, probably with an increasing impact on
Treasury bond yield and they will try to avoid fire sales of commercial loans as they are illiquid
assets. The central bank would probably sell treasury bonds to accommodate lower banking
reserves and therefore there is a moderate reduction of seigniorage. The balance sheet of
commercial banks and the central bank may look like this in period t=1a:
Commercial Bank (t=1a)
Assets
2550 Commercial Loans

Liabilities
2800 Individuals’ Demand Deposits

110 Treasury Bonds
140 Reserves at Central Bank

Connection to Pix is mandatory for payment service providers (PSPs) with more than half a million active customer
accounts and voluntary for all other BCB-regulated banks and non-bank entities. Pix interoperability enables
individuals, businesses, and government entities to perform instant payments and transfers. For their part, the
Reserve Bank of India (RBI) and India’s national bank association established a new entity called the National
Payment Corporation of India (NPCI), which developed an instantaneous payment interface called the United
Payments Interface (UPI) in 2016. Like in Brazil, RBI regulates UPI.

INTERNATIONAL MONETARY FUND

16

IMF WORKING PAPERS

Digital Money and Central Banks Balance Sheets

Central Bank (t=1a)
Assets
1140 Interest Bearing Financial Assets

Liabilities
1000 CiC
140 reserves

As we can see, base money declines from 1150 to 1140 and traditional money (M1) from 4000 to
3800. If stablecoins are included in the definition of broad money, there is no change, and the
figure is 4000. Given the nature of the stablecoins, it is likely to have relative more use as a
means of payment and less as store of value compared with the average banking demand
deposits (in particular time deposits) and therefore the implicit velocity from the quantitative
money equation will be larger. 14 An efficient liquidity management for the Fintech will depend
on the depth and liquidity of the treasury bond market (including repo market).
In this example, the stablecoins are out of the regulatory (and counterparty) perimeter of the
central bank. Under the current framework, central banks run OMO (open market operations)
with depository institutions (mainly banks) because they are in charge of the functioning of the
payment system. Central banks generally do not transact with capital markets (i.e., mutual funds,
hedge funds, private equity, etc.), except in extraordinary circumstances to protect the
transmission mechanisms of monetary policy. An open question in this context if authorized
stablecoins and PSP should have access to open current accounts at the central bank. 15
Following our example, in the long run commercial banks may be willing to reaccommodate
their portfolio as before given that loans will start to be repaid. Assume that this event will take
place in period t=2a. Now the balance sheet of commercial banks will look like this (we assume
no changes in central bank’s balance sheet).

Fisher’s quantitative money theory (MV=PT), suggests that money (M) times velocity of circulation (V), equals
average price level (P) times number of transactions (T).
15
The impact of stablecoins (backed by central bank reserves) on credit and monetary aggregates is similar to the
case of CBDC replacing deposits (Case vi). The reason is that in both cases money base increases and banking
multiplier falls.
14

INTERNATIONAL MONETARY FUND

17

IMF WORKING PAPERS

Digital Money and Central Banks Balance Sheets

Commercial Bank (t=2a)
Assets
2380 Commercial Loans

Liabilities
2800 Individuals’ Demand Deposits

280 Treasury Bonds
140 Reserves at Central Bank

Banking credit to the private sector falls 4.8 percent. Large companies may be able to substitute
the reduction of the supply of commercial loans by issuing corporate debt in the capital market or
from external bank lending. In the case of small and medium-sized enterprises (SMEs) that will
not be the case. However, there are other developments in the Fintech world regarding the credit
market that could improve financial conditions to SME and consumer loans. On the transmission
mechanism of monetary policy, as banking credit shrinks in this case, credit will be lower.
Case (iii) Substitution of Bank Deposits by Stablecoins in Foreign Currency in a Partially
Dollarized Economy
It is envisaged that the impact on stablecoins in foreign currency would be larger in countries
with relatively high inflation and existing capital controls and restrictions to open deposits in
foreign currency. The reason is that there are countries with currencies with the potential to be
partially replaced by foreign currency, but regulations make it impossible. To illustrate the
argument let’s assume now that it is allowed to open dollar deposits in period t=0 and the
regulation establish a reserve requirement or liquidity ratio of 25 percent for those deposits. 16 At
period t=0 the balance sheets are:

Armas et al. (2015) makes the case of high reserve requirements for foreign currency liabilities as a case of lack of
lender of last resort (LOLR) in foreign currency. We asume a zero foreign currency exposure for banks (i.e., no long
or short positions).
16

INTERNATIONAL MONETARY FUND

18

IMF WORKING PAPERS

Digital Money and Central Banks Balance Sheets

Commercial Bank with Dollar Deposits (t=0)
Assets

Liabilities

2380 Commercial Loans in Domestic Currency

2800 Individuals’ Demand Deposits in Local
Currency

280 Treasury Bonds in Domestic Currency

200 Individuals’ Demand Deposits in Foreign
Currency

140 Reserves at Central Bank
150 Commercial Loans in Foreign Currency
50 U.S. Treasury Bond

Central Bank in the Case of Partial Dollarization (t=0)
Assets
1140 Interest Bearing Financial Assets

Liabilities
1000 CiC
140 Reserves

Now in period t=1 there are stablecoins in foreign currency, but the demand for them will be
lower because citizens have the option to keep dollar savings in the local banking system. For the
same reason, the substitution between stable coins and deposits in foreign currency will be larger
than the one with local deposits. Assuming the latter is zero and local credit in dollars cannot
change in the very short term, the balance sheets are the following:
Fintech Issuer of a Stablecoin in Foreign Currency when Dollar Deposits are Allowed (t=1)
Assets

Liabilities

100 Infrastructures

100 Capital

100 U.S. Treasury Bond

100 Stablecoins in Dollar

INTERNATIONAL MONETARY FUND

19

IMF WORKING PAPERS

Digital Money and Central Banks Balance Sheets

Commercial Bank with Dollar Deposits (t=1)
Assets

Liabilities

2380 Commercial Loans in Domestic Currency

2800 Individuals’ Demand Deposits in Local
Currency

280 Treasury Bonds in Domestic Currency

100 Individuals’ Demand Deposits in Foreign
Currency

140 Reserves at Central Bank

75 External Borrowing or, CB liquidity to support
foreign currency liquidity (e.g., FX swap line)

150 Commercial Loans in Foreign Currency
25 U.S. Treasury Bond

Central Bank in the Case of Partial Dollarization (t=1)
Assets
1140 Interest Bearing Financial Assets and CB
Liquidity Support in Foreign Currency to
Commercial Banks

Liabilities
1000 CiC

140 Reserves

Capital outflows bring shortage of dollar liquidity in the banking system, and it is faced by
borrowing from international commercial banks; or some liquidity support in foreign currency
from the CB (as it takes time to adjust and cushion from such capital movements).
In period t=2 when credit can change, and commercial banks would reshuffle their portfolio to
overcome the liquidity shortage (i.e., the external borrowing of 75 or, liquidity support, will no
longer be needed), with a lower amount of commercial banks in dollars. In our example, the
balance sheets are:

INTERNATIONAL MONETARY FUND

20

IMF WORKING PAPERS

Digital Money and Central Banks Balance Sheets

Commercial Bank with Dollar Deposits (t=2)
Assets

Liabilities

2380 Commercial Loans in Domestic Currency

2800 Individuals’ Demand Deposits in Local
Currency

280 Treasury Bonds in Domestic Currency

100 Individuals’ Demand Deposits in Foreign
Currency

140 Reserves at Central Bank
75 Commercial Loans in Foreign Currency
25 U.S. Treasury Bond

Central Bank in the Case of Partial Dollarization (t=2)
Assets
1140 Interest Bearing Financial Assets

Liabilities
1000 CiC
140 Reserves

The example shows that the impact on banking credit for a partial dollarization is lower because
local deposits in foreign currency allow banking lending in the same currency. 17
Case (iv) Substitution of CiC by E-money
As shown in Kahn et al. (2022), there is a declining trend of CiC before the pandemic and as the
world is leaving the COVID-19 era and entering to a “new” normal situation, the development of
digital payment tools will likely continue driving this trend. Let us see now this case starting
from period t=0 (without partial dollarization of deposits) and in period t=1b e-money surges
induces a substitution of CiC.

17

A fixed exchange regime will be more vulnerable than a floating rate regime.

INTERNATIONAL MONETARY FUND

21

IMF WORKING PAPERS

Digital Money and Central Banks Balance Sheets

Commercial Bank (t=1b)
Assets

Liabilities

2720 Commercial Loans

3000 Individuals’ Demand Deposits

320 Treasury Bonds

200 Phone Company Demand Deposits

160 Reserves at Central Bank

Central Bank (t=1b)
Assets
960 Interest Bearing Financial Assets

Liabilities
800 CiC
160 Reserves

The substitution of CiC by e-money is equivalent to a reduction of preference of currency. This
would entail a larger banking multiplier effect and lower money base (1150 to 960), keeping
constant the total amount of Money (4000); CiC as a fraction of M1 is 20 percent now. As a
consequence, the supply of credit to private sector increases and the seigniorage falls.
Case (v) Substitution of CiC by CBDC
Today, money is mainly issued by banks, as share of CiC is relatively low, in particular in
developed economies (with some exceptions such as Switzerland). In developing economies, the
size of CiC as share of GDP is relatively high as many citizens do not have access to banking
account (although there is a declining trend). According to Findex 2021, the access of the adult
population to at least one banking account is almost universal in Latin America. However, the
heterogeneity is large within the region, where Brazil has a relative high share of people with a
banking account (84 percent) and Nicaragua only 26 percent. These figures may explain the fact
that financial inclusion is a more important argument for emerging and developing economies in
the assessment of issuing CBDC.
Next, let us see the case when a CBDC replaces CiC (e.g., for financial inclusion for instance).
The only change here is the composition of base money in period t= 1c. Therefore, there is no
impact on the transmission mechanism of monetary policy, there is a clear improvement of the
efficiency of the payment system and no changes in seigniorage. Digital money velocity is higher

INTERNATIONAL MONETARY FUND

22

IMF WORKING PAPERS

Digital Money and Central Banks Balance Sheets

(Kahn et al., 2022). Thus CiC (t1)+ CBDC (t1) may/may not be less than CiC(t0), so seigniorage
from base money may decline.
Regarding monetary liquidity management, central banks should now include daily movements
of CBDC demand as a new autonomous factor to estimate the magnitudes of OMO needed to
reach operational target (policy rate or a banking reserve target).
Central Bank (t=1c)
Assets

Liabilities

1150 Interest Bearing Financial Assets

800 CiC
200 CBDC
150 Reserves

Case (vi) Substitution of Bank Deposits by CBDC
In this case where CBDC gets enough traction to replace demand deposits, assuming that the
services (and underlying economics) of the CBDC exceeds private banking services. The central
bank’s balance sheet increases (from 1150 to 1340 in our example) and there is a case of
financial disintermediation. This scenario has a low probability according to a survey of central
banks (BIS, 2022).
Let us see the balance sheets in period t= 1d, when it is not possible to reduce the stock of credit.
Commercial banks need to sell treasury bonds to face this liquidity shortage and the central bank
must inject liquidity (may be purchasing treasury bonds) to accommodate the higher base money
demand.
Commercial Bank (t=1d)
Assets
2550 Commercial Loans

Liabilities
2800 Individuals’ Demand Deposits

110 Treasury Bonds
140 Reserves at Central Bank

INTERNATIONAL MONETARY FUND

23

IMF WORKING PAPERS

Digital Money and Central Banks Balance Sheets

Central Bank (t=1d)
Assets

Liabilities

1340 Interest Bearing Financial Assets

1000 CiC
200 CBDC
140 Reserves

In our numerical example commercial banks can face the liquidity shortage because they have
enough good collateral (treasury bonds with a deep secondary market and low credit risk). One
obvious buyer of those treasury bonds is the central bank as base money demand has increased.
However, in an extreme case the stock of holdings of treasury bonds or other similar assets may
be lower than the amount needed to cover banking withdrawals. In this extreme case, liquidity
shortage may increase interbank rates significantly above the policy rate if the central bank does
not extend the list of collateral to run monetary operations.
In period t=2d commercial banks reduce the amount of commercial loans from 2550 to 2380 to
match the availability of deposits. Therefore, we expect a weaker interest rate and credit channel
of monetary policy.
Commercial Bank (t=2d)
Assets
2380 Commercial Loans

Liabilities
2800 Individuals’ Demand Deposits

280 Treasury Bonds
140 Reserves at Central Bank

VI. Conclusion
Digital money will have implications for banking intermediation, external capital flows, and
central banks balance sheet. If the digital money revolution occurs under a fiscal dominance
regime and reduces seigniorage, central banks may be inclined to take actions and compromise
that situation. Under a monetary dominance regime (i.e., independent monetary policy), the
development of a payment system should focus on how to improve its efficiency, without
jeopardizing the monetary policy transmission.
INTERNATIONAL MONETARY FUND

24

IMF WORKING PAPERS

Digital Money and Central Banks Balance Sheets

Depending on the type of substitution that takes place between traditional money and digital
money, there could be banking disintermediation as deposits move away from commercial
banks. Stable coins in foreign currency may increase international financial integration, while
developing economies will face new challenges vis-à-vis capital flows. Furthermore, the
operational aspects of digital money and the impact of CBDC on financial disintermediation are
in their infancy; thus, it may be preferable to use interest rate on CDBC only in extraordinary
cases (and this remains a debatable issue as policy makers implement their pilot studies in this
area).
From a monetary policy perspective, this paper summarizes cases where the interest rate and
credit channel may become weaker when there are instances such as Fintech issues stablecoins
being backed by treasury bonds, or there is substitution of bank deposits by CBDC or even when
there is substitution of bank deposits by stablecoins in foreign currency—a possible scenario in
developing economies. The common factor in these cases is a reduction of banking credit.
However, when CiC is substituted by private digital money that is deposited in commercial
banks, there may be an expansion of banking credit.

INTERNATIONAL MONETARY FUND

25

IMF WORKING PAPERS

Digital Money and Central Banks Balance Sheets

References
Adrian, Tobias, and Tommaso Mancini-Griffoli, (2019), “The Rise of Digital Money,” July.
Armas, Adrian, Alan Ize, and Eduardo Levy-Yeyati, editors (2006). “Financial Dollarization:
The Policy Agenda.” International Monetary Fund. Palgrave Mac Millan.
Armelius, H., Paola Boel, Carl Andreas, and Marianne Nessen (2018). “The E-krona and the
Macroeconomy.” Sveriges Riksbank Economic Review.
Baumol, William J., (1952). The Transactions Demand for Cash, Quarterly Journal of
Economics, 66, 545-556.
Bernanke, Ben, and Alan Blinder (1988). “Credit, Money and Aggregate Demand.” American
Economic Review, American Economic Association, Vol 78 (2), pages 435-439, May.
Bindseil, Ulrich (2016). Jackson Hole Speech, “Evaluating Monetary Policy Operational
Frameworks.” August 26.
BIS (2022). “CBDC in Emerging Market Economies.” Papers 123. Monetary and Economic
Department.
Brunnermeier, Markus, and Jonathan Payne (2022). “Platforms, Tokens, and Interoperability.”
Princeton, June.
Eichengreen, Barry, “Golden Fetters: The Gold Standard and the Great Depression 1919-1929.”
Oxford Scholarship. Online, 1993.
Evans, D., and R. Schmalensee (2016). “Matchmakers: The New Economics of Multisided
Platforms.” Harvard Business Review Press.
Gourinchas, Pierre, Helen Rey, and Maxime Sauzer (2019). “The international Monetary and
Financial System.” Annual Review.
International Monetary Fund (2019), “The Rise of Digital Money: A Strategic Plan to Continue
Delivering on the IMF’s Mandate,” July 29.
_____________________(2020), Digital Money Across Borders: Macro-Financial Implications,
October 29.

INTERNATIONAL MONETARY FUND

26

IMF WORKING PAPERS

Digital Money and Central Banks Balance Sheets

Kahn, Charles, and Manmohan Singh (2021), “If stablecoins are money, they should be backed
by reserves – Risk.net.”
Kahn, Charles, Manmohan Singh, and Jihad Alwazir (2022). “Digital Money and Central Bank
Operations.” IMF Working Paper.
Leeper, Eric (1991). Equilibria Under Active and Passive Monetary and Fiscal Policies, Journal
of Monetary Economics, 17, 129-147.
Mancini Griffoli, Tommaso, Maria Soledad Martinez Peria, Itai Agur, Anil Ari, John Kiff, Adina
Popescu, and Celine Rochon, (2017), “Casting Light on Central Bank Digital Currency.”
Reserve Bank of India (2021). “Central Bank Digital Currency—Is this the Future of Money?”
T Rabi Shankar, Deputy Governor of RBI, July 22.
Schwartz, Anna. “Banking School Currency School, Free Banking School” in Eatwell, Milgate
and Newman (2018) The New Palgrave-Macmillan, pages 694-695.
World Bank (2022). “The Global Findex Database 2021: Financial Inclusion, Digital Payments,
and Resilience in the Age of COVID-19.”

INTERNATIONAL MONETARY FUND

27

IMF WORKING PAPERS

Digital Money and Central Banks Balance Sheets

Annex I. Potential Impact of Value-Based CBDC
with a Cap for Relative Small Amounts
Country Case/
Function of Money

Unit of
Account

Mean of Transactions

Store of Value

Issuer of
International
Currency

None.

CBDC may have relatively small scope
to be demanded by residents as those
economies already use non-cash tools
for most of transactions. More
substitution possibilities with CiC in
particular from non-residents holders.

Minor impact in the case of
residents. More demand from
non-residents as safe-haven
asset, in particular those with
low access to foreign currency
accounts.

Open Economy
Under Dominant
Currency
Paradigm

None.

CBDC may encourage competition for
faster and lower transaction cost with
traditional banking payment system in
domestic currency. It has a potential as
financial inclusion mechanism if
traditional banking is not able to fill the
gap; more substitution with CiC is
expected.

Minor impact, except in periods
of banking crisis where demand
for both CiC and CBDC may
increase.

Partially and
Highly Dollarized
Economy

None.

CBDC may encourage competition for
faster and lower transaction cost with
traditional banking payment system in
domestic currency. It has a potential as
financial inclusion mechanism if
traditional banking is not able to fill the
gap. It is expected more substitution
with CiC. It could reduce payments
dollarization.

Minor impact, except in periods
of banking crisis where demand
for both CiC and CBDC may
increase.

Note: While this may be a possibility in the future, all currently ongoing retail CBDC initiatives center on domestic access
only (this is different for wholesale CBDC).