Monetary Policy, Monetary Areas, and Financial Development with Electronic Money

WP/04/122


Monetary Policy, Monetary Areas,
and Financial Development
with Electronic Money

Marco Arnone and Luca Bandiera

---

© 2004 International Monetary Fund
WP/04/122

IMF Working Paper

Monetary and Financial Systems Department

Monetary Policy, Monetary Areas, and Financial Development
with Electronic Money

Prepared by Marco Arnone and Luca Bandiera1

Authorized for distribution by Piero Ugolini

July 2004

Abstract

This Working Paper should not be reported as representing the views of the IMF.
The views expressed in this Working Paper are those of the author(s) and do not necessarily represent
those of the IMF or IMF policy. Working Papers describe research in progress by the author(s) and are
published to elicit comments and to further debate.

Electronic money (e-money), as a network good, could become an important form of
currency in the future. Such a development could affect monetary policy effectiveness. If an
increased use of e-money substantially limits the demand for central bank reserves, this
limitation would require changes in the central bank operational target and a closer
coordination of monetary and fiscal policies. Also, the optimal size of monetary unions
would be different. However, the current level of e-money use does not seem to pose a threat
to the stability of the financial system. Thus, central banks can successfully implement the
objectives of monetary policy.

JEL Classification Numbers: E40, E42, E50, F33

Keywords: electronic money, monetary policy, monetary unions.

Author’s E-Mail Address: MArnone@imf.org, LBandiera@worldbank.org



1 Luca Bandiera is a consultant at the World Bank. The paper was written while Mr. Bandiera
was affiliated with the Kiel Institute for World Economics. The authors would like to thank
Rodolfo Maino, Andrea Schaechter, Martin Sommer, Piero Ugolini, and participants at a
Kiel Institute for World Economics seminar for their helpful comments.

---

- 2 -



Contents
Page

I. Introduction .......................................................................................................................3
II. Definition and Classification of E-Money ........................................................................4
III. E-Money and Network Externalities ................................................................................6
IV. Payment and Settlement .................................................................................................10
V. Financial Development Issues ........................................................................................11
VI. The Transmission of Monetary Policy ...........................................................................13
VII. E-Money and Monetary Areas .......................................................................................18
VIII. Conclusions....................................................................................................................21

References................................................................................................................................40

Tables
1. System Design Features...................................................................................................23
2. Data on Use of E-Money .................................................................................................33
3. Characteristics and Utilization of E-Money Products, 2000–2003 .................................38
4. Changes in Characteristics and Utilization of E-Money Products, 2000–2003...............39

Figures
1. Electronic Money Schemes................................................................................................5
2. The Fulfilled Expectation Demand Curve .........................................................................8
3. The Market for Overnight Balances in a Cashless Society..............................................16
4. Optimal Size of Networks................................................................................................19
5. The Effect of Information Technology on the Optimal Size of Monetary Unions..........21

---

- 3 -



INTRODUCTION
The recent evolution of the technology for financial transactions poses interesting questions
for policymakers and financial institutions regarding the suitability of the current institutional
arrangements and the availability of instruments to guarantee financial stability, efficiency,
and effectiveness of monetary policy. The aim of this paper is to dispel some of the most
extreme predictions.

The disappearance of money, central banks made redundant, monetary policy made
irrelevant are not logical impossibilities, but extreme events. It is not impossible to envision a
world with these most extreme characteristics (King, 1999) and to describe its functioning in
the new setup, but how likely are these events to occur? The view that they will happen in the
foreseeable future cannot be shared. From an operational point of view, we need to ask
whether the actual institutional arrangements can respond properly to the recent wave of
financial infrastructure innovations and if their policy instruments are still viable.

Regulatory concerns are raised by internet banking, electronic finance (e-finance), and
e-money but are not addressed in these notes. Here the focus is primarily on the forces
sustaining the development of e-money and on the central bank ability to conduct monetary
policy in the presence of e-money.2

This paper argues that e-money, as a network good, could become an important form of
currency in the future. Such a development would influence the effectiveness and
implementation of monetary policy. If an increased use of e-money substantially limits
demand for central bank reserves, it would require changes in the operational target of the
central bank and a closer coordination of monetary and fiscal policies. The optimal size of
monetary unions could differ should e-money play a prominent role.

However, the current level of e-money use does not pose a threat to the stability of the
financial system, and central banks have the means to successfully implement the objectives
of monetary policy.

With respect to the current debate on the consequences of the introduction of e-money on the
implementation of monetary policy and in contrast with earlier work on this topic (Freedman,
2000; Friedman, 2000; Goodhart, 2000; and Woodford 2000), this paper argues that central
banks can lose control over monetary policy if the government does not run a responsible
fiscal policy. Central banks could react to this threat by introducing additional regulation or
by resizing the monetary areas which they ought to regulate, as we shall argue in later
sections.

This paper is structured as follows. Sections II and III describe e-money and its
characteristics as a network good. Section IV analyzes payments and settlement issues.

2 Details about regulations can be found in ECB (1998), and BIS (2000 and 2001a).

---

- 4 -


Financial development issues are discussed in Section V, while Section VI describes the
transmission mechanism of monetary policy in the presence of e-money. Section VII
redefines the optimal size of monetary areas, and Section VIII concludes.


II. DEFINITION AND CLASSIFICATION OF E-MONEY
E-money is defined as an “electronic store of monetary value on a technical device that may
be widely used for making payments to undertakings other than the issuer without
necessarily involving bank accounts in the transaction, but acting as a prepaid bearer
instrument” (ECB, 1998). There are three main types of e-money: e-cash, network money,
and access products. E-cash includes reloadable electronic purses and multi-purpose stored
value cards.3 Network money defines funds stored in software products that can be used for
payments or transfers over communication networks (i.e., the Internet). Access products
allow agents to access their bank accounts and transfer funds (Prinz, 1999). These latter
products are new ways of executing transactions with existing forms of money (OECD,
2002).

E-money schemes have some distinguishing features: (i) technical implementation requires
computer and software expertise; (ii) these products are liabilities on the balance sheet of the
issuer;4 (iii) transferability is somehow limited: while relatively straightforward from
consumer to merchant, to be redeemed by the issuer, it is unusual from consumer to
consumer (peer-to-peer); and (iv) transactions can be easily recorded (no anonymity).5

Some schemes allow spending e-money on a single good or service, others on a set of goods
and services (multi-purpose e-money). Technology clearly leans in the direction of
expanding the set of goods and services that agents can purchase within the same scheme.
Schemes can involve two parties, where the issuer of e-money is also the provider of goods
and services; or schemes can involve three parties, where the issuer and provider functions
are separate.


3 As distinct from credit cards, debit cards, automated teller machine (ATM) cards and the
like, which do not have any intrinsic value themselves but allow customers to transfer funds
between bank accounts, and single-purpose payment cards which are essentially an
acknowledgement of a pre-payment (Hawkins, as found in BIS, 2001b).
4 Issuers should be appropriately supervised to avoid excess supply that could threaten
market stability (ECB, 1998).
5 Prinz (1999) considers electronic payment systems where anonymity of the consumer is
guaranteed by the credit or debit transaction system supervised by the issuer. However, these
systems do not preserve the complete anonymity of the buyers because the issuer controls the
inherent creditworthiness.

---

- 5 -


There are several characteristics that distinguish e-money from traditional bank notes. The
latter are infinitely transferable and completely anonymous, totally risk-free from a
settlement point of view, and can be used for any sort of transaction without restrictions,
since bank notes are legal tender. Traditional bank notes have the maximum benefit from
positive network externalities, as we shall see in Section III. E-money is not a liability
necessarily settled on the books of the central bank. For this reason e-money exploits
network externalities to a much more limited extent, depending on the size of the scheme
(number of goods and services offered, number of people using the scheme, ease of access,
and so forth).

In assessing e-money as a network good (see Section III) vis-à-vis government money, the
following elements could also be considered: (i) transaction costs tend to be higher compared
with traditional bank notes, as e-money requires specific technologies; (ii) the potential
effects of taxation could include an ex ante reduction on the use of private money;
(iii) collection of seigniorage (or other forms of revenues) would be internalized by the
issuing network; (iv) interest sensitivity to other forms of interest-bearing assets in the
context of a multiplicity of competing monies and asset substitutability; and (v) expectations
and risks regarding alternative monies.

The Bank for International Settlements (BIS) regularly surveys the development of e-money.
Since 2000, annual statistics have been compiled (Tables 1 and 2). The last survey covers the
period from 2002 to 2003. Even though the time period covered by the surveys is limited, a
few general conclusions can be drawn on current trends in e-money developments. E-money
products are generally card-based; however, the network-based systems’ growth rate has
been 100 percent and 67 percent in 2001 and 2002, respectively, compared to 27 percent and
36 percent, respectively, in card-based systems (Figure 1). They can be used to execute a
multiplicity of payments, and most of them are adapted for network payments. The direct
transferability of funds among users is limited, as is the possibility to use e-money for
international transactions (Tables 3 and 4).

Figure 1. Electronic Money Schemes
60
50
ems 40
st
f
Sy
E-cash
30
Network money
mber o 20
u
N 10
0
2000
2001
2002

Sources: Bank for International Settlements (2000, 2001a, and 2004); and own calculations.

---

- 6 -


Since 2000, the number of e-money systems has increased steadily worldwide, but Asia and
Africa have experienced the fastest increase. In 2003, about 40 percent of e-money systems
were located in Western Europe, down from 50 percent in 2001. New systems were
introduced in Italy, Norway, and Finland. In the United Kingdom, all existing systems were
discontinued by end-2001. In Central and Eastern Europe new systems were introduced in
Russia and in the Czech Republic in 2002.6 In North America, the use of e-money systems
remains limited because credit cards are traditionally used to execute even small-value
transactions. In Central and South America and the Caribbean, e-money systems had an early
start, but many proved to be unsuccessful. Since 2000, all systems in Mexico, Costa Rica,
Honduras, and Venezuela were discontinued. In 2002, a new system was introduced in
Brazil. E-money products are particularly widespread in Asia, where 21 new systems were
launched in India, South Korea, Malaysia, the Philippines, and Singapore. In Hong Kong
SAR, the existing system has considerably enlarged the number of users. In Africa, eight new
systems have been created, and none of the previously existing systems has failed.

Western Europe is the most mature market for e-money systems, with the lowest turnover
rate and the largest volume of purchases. However, the number of card issued is decreasing,
while the purchase volume is increasing at a very high pace. Systems existing since 2001
have considerably reduced the number of cards issued, but about 95 percent of total
purchases are still carried out through these systems. Therefore, fewer end-users use e-money
systems more frequently. In Asia, the number e-cards issued is increasing at a higher pace
than the volume of purchases. In contrast to Western Europe, many new systems were
recently launched in Asia and the number of e-cards does not correspond to the number of
users that carry out transactions.

From 2000 to 2003, the value of the average purchase has decreased from US$4.7 to US$2.3.
The Czech Republic and Brazil represent the only notable exceptions. This confirms the
belief that e-money, as opposed to credit and debit cards, will be mainly used for small-size
payments.


III. E-MONEY AND NETWORK EXTERNALITIES
E-money is a typical network good and, as such, the impact of its introduction as a means of
exchange can be analyzed through the concepts of network and the theory of network
externalities. E-money, as a network good, could achieve a diffusion deemed sufficient to
substitute for traditional money. However, compatibility, efficiency, and regulatory issues
may require the intervention of national regulators for e-money to be efficiently used in the
monetary system.

Network externalities consist of the utility increase for network members derived from the
increase in the number of network users (Katz and Shapiro, 1985; Van Hove, 1999). The
usage of a specific good “called network good” characterizes the specific network, (e.g.,

6 Including Turkey.

---

- 7 -


telephone, fax, computer connected to Internet). In the case under consideration, network
externalities are usually positive (although limited capacity with rationing––like queuing, or
limited access––is a typical example of a negative externality) and arise on the demand side.

Generally speaking, the market value of the network good is inherently related to its
connection to the network: a network good has limited or no value outside the network.
Moreover, the scope of the network externality is limited to the network itself, that is: no
externality arises outside the network, so that only the members of the network can benefit
from the externality generated by the additional member.7

What network size will agents consider sufficient to start using e-money for transactions?
The concept of critical mass point helps answer the question. A market for network goods
will not exist unless the expected size is greater than a minimum level (Economides and
Himmelberg, 1995). 8 The price of a network good increases with the expected size of the
network and decreases with the number of participants (Economides, 1996). Abstracting
from risk considerations, the demand for network goods is downward sloping for each level
of expected size. In Figure 2 (Van Hove, 1999, based on Economides and Himmelberg,
1995), the effect of network externalities is represented by the series of inverse demand
curves p(n, nei). The price p is a function of the network participants n and of the expected
dimension of the network nei, where the total size of the population of the economy is
normalized to 1.

In equilibrium, the level of network demand matches its expected dimension. Therefore, it is
possible to construct a fulfilled expectation demand curve, joining all equilibrium points on
the demand schedules. The fulfilled expectation demand curve is represented by the bold
curve p(n, n). The demand schedule for network goods includes the y-axis up to the origin,
since e-money, as a network good, does not have any stand-alone value and nobody would
like to keep e-money in the event that no network is expected. Under a general set of
conditions,9 it is possible to show that the fulfilled expectation demand curve is single-
peaked and that the price that the last consumer is willing to pay (n = 1) is necessarily zero
because there are no further gains in utility from joining a network as big as the whole
population in the economy (Economides and Himmelberg, 1995).


7 However, the value of the network and, consequently, the value of the network good
depend on the probability that customers switch from one network to another. This means
that the network externality might have an indirect impact on another network.
8 For example, nobody will buy a telephone if he or she is the only owner.
9 The inverse demand curve is derived from a utility function that depends on the stand-alone
value of the good and that is increasing, at a decreasing rate, with respect to the number of
network participants (Economides and Himmelberg, 1995).

---

- 8 -


Figure 2. The Fulfilled Expectation Demand Curve


c,p
c0
p(n, n)
p(n, ne2)
p(n, ne1)
0
ne1 ne2
n0
1
n


Furthermore, the equilibrium demand schedule is increasing for small n if one of the
following conditions holds: (i) the utility of every consumer in a network of zero is zero;
(ii) there are immediate and large external benefits to network expansion for very small
networks; and (iii) there is a significant density of high willingness-to-pay consumers who
are just indifferent on joining a network of approximately zero size (Economides, 1996).

In the case of perfect competition, the equilibrium in the market for the network good
presents a critical mass point, corresponding to the minimal non-zero equilibrium size of the
network. This point is represented by n0 in the diagram. For marginal costs (MC) greater than
c0 the optimal size of the network is zero, while for MC smaller than c0 multiple equilibria
exist: zero network size and the points on the x-axis correspondent to the first and the last
intersection of the MC with the demand schedule. Only the outer intersection is a stable
Pareto-efficient equilibrium (Van Hove, 1999).

Does e-money present the characteristics considered sufficient to reach a critical mass point?
Realistically, e-money has no value as a stand-alone good (condition i), if it is not used for
transactions. The demand for e-money can be justified by the relevance of new forms of
commerce and financial transactions (e-commerce and e-finance), which are carried out in
virtual markets, thanks to advances in information technology (IT) (Prinz, 1999). Given
technological advances and the functions satisfied by different forms of money, the decrease
in the cost of use of e-money could generate enough demand to reach a critical mass
(conditions ii and iii).

Even though e-money could generate an initial equilibrium while its network grows larger,
the issue of e-money under perfect competition is inefficient because the marginal social
benefit of network expansion is greater than the benefit that accrues to the single participant.
Prinz (1999) shows that, under perfect competition, the demand for e-money is less than
socially optimal.

---

- 9 -


Economides (1996) demonstrates that the presence of a single private issuer of e-money
leads to an even smaller network than in the case of perfect competition.

In the case of oligopoly, Economides (1996) shows that a social optimum could be reached if
issuers adopt a compatible technology that supports the networks. Issuers will agree to
cooperate if the profits deriving from a bigger network are not offset by the increased level of
competition in the market for e-money. As a result, the chances to develop sustainable
networks depend on the balances of multiple conditions.

According to Van Hove (1999), e-money networks must satisfy six conditions to develop.
First, production of e-money must present increasing economies of scale, so that few large
issuers will constitute the market. Second, network effects must be strong for consumers to
pay higher prices for larger networks. Third, network effects must not weaken too quickly;
thus small incompatible networks are not economically sustainable. Fourth, e-money must
offer uniform services in order to reinforce the importance of the network dimension for the
utility of consumers. Fifth, the installed base must be fairly visible, so that consumers are
able to choose which network to join on the basis of their expectations. And sixth, the
network must have a minimum geographical scale to avoid the possibility of excessive
supply of local e-money.

The first condition refers to technological constraints and the second to the demand for
network products. The remaining conditions assure that neither consumers nor issuers
demand or produce too many incompatible forms of e-money. The numerous conditions that
e-money must fulfill to create the sufficient size of the network to be accepted as a medium
of exchange and store of value are a delicate balance between compatibility and competition.

Each form of e-money competes with other forms of e-money and with traditional money,
which, as legal tender, maximizes network benefits. Competition limits the diffusion of
e-money because networks will not reach socially optimal sizes and more issuers are less
likely to agree upon technological standards. Notwithstanding all of these limitations,
demand for e-money could increase with the size of new forms of e-markets for goods and
services. The problem of compatibility of e-money argues for the intervention of national
regulators (Prinz, 1999): central banks and financial sector regulators could address the
compatibility problem by making e-money legal tender and supervising its issuers.

Alternatively, e-money could develop efficient network sizes only in a limited number of
sectors of the economy. Even in this scenario, the subsequent reduction of the use of
traditional banks’ money might have consequences for the functioning of the payment and
settlement system and the conduct of monetary policy, as we shall see in Sections IV and VI.

---