1. On October 27, 2000, the EMI directive on the taking up, pursuit and prudential supervision of the business of electronic money institutions was promulgated. This directive, to be implemented by April 27, 2002 at the latest, allows organisations other than credit institutions to issue electronic money. Such organisations, referred to as electronic money institutions, need to be properly supervised, but they can operate in all the European countries under a single license. Considering the wider context of the rapidly evolving electronic commerce business, the directive explicitly states the importance of a regulatory framework that avoids hampering technological innovation and that harmonises the prudential supervision of electronic money.
2. In October 2002, 1.1a2, the Association for E-money Institutions in the Netherlands, conducted an investigation into the implementation status of the EMI directive. The results show that:
- five of the fifteen member states have not yet implemented the directive,
- two substantially diverging supervisory approaches are applied.
3. We conclude that, at present, there is insufficient harmonisation of the prudential supervision framework for electronic money. More specifically, the concept of a technology-neutral approach has not yet been fully adopted and harmonised. As a result:
- a particular e-money system may qualify as e-money in one member state, while in another member state it will be considered a remote banking system,
- even within member states, functionally similar e-money systems may qualify differently on the basis of their distinct technical features.
4. Since September 2000, market players have based their investments on the expectation that as of 2003 a harmonised legal framework for e-money institutions in Europe would be in place (regardless of the technical choices made). These market players now face the problem that their future market position and market potential may be unduly limited by the approach taken by their local supervisor. The chance exists that in some countries local supervisors may start compliance activities based on an ill-understood concept of e-money, the market of e-money institutions or the technology applied.
5. The members of 1.1a2 wish to prevent a situation from arising where some supervisors in Europe require market players to execute changes in organisation, product or technology, that in retrospect and in comparison with other member states turn out to be unduly restrictive. We note that the implementation of the EMI directive is already facing delay due to the late implementation in five member states. As the establishment of a consistent European level playing field for e-money issuers is already taking more time than expected, we recommend that the competent supervisors in the European member states not rush forward with local approaches to the supervision of electronic money institutions, but that they agree on a harmonised prudential supervision regime for electronic money.
6. In summary, 1.1a2 urges European competent supervisors:
1- to postpone local compliance activities in the domain of electronic money and electronic money institutions until their supervisory approaches have been harmonised and are in accordance with the principle of technological neutrality,
2- to consider the functional approach of the Financial Services Authority as the basis for the harmonised approach to supervision of electronic money institutions,
3- to recognise the relevance of recent and future product developments in the mobile telecom industry to the electronic money debate.
Introduction and structure of the position paper This position paper discusses the concept of electronic money, its market potential and the current regulatory issues. 1.1a2, the Association for E-money Institutions in the Netherlands, publishes this paper at this specific moment in time as substantial differences appear to exist in the way the EMI directive1 is understood and implemented by regulators/supervisors in the various member states of the European Union. The consequences of this divergence are best illustrated by an example. In the example we take a detailed look at three pre-paid electronic purse systems, all of which make use of electronic purse, chipcard and mobile phone technology.
A. Dual slot e-purse system
In this first system, the consumer uses a regular bank chipcard that has an electronic purse application. The consumer enters the chipcard into his/her mobile phone that is equipped with a second slot to read the IC chip with the e-purse application. The mobile phone operates here as the payment terminal that facilitates the communication between the e-purse-application on the chipcard and the Security Application Module (SAM) of the merchant. This SAM validates and accepts the e-purse payments and may reside on a remote computer server of the merchant.
B. E-purse in Subscribe Identification Module (SIM)
In the second set-up, banks and mobile operators have agreed not to equip mobile phones with chipcard readers but to use a part of the consumer’s SIM to store the electronic purse application. As a result, the e-purse software must be loaded into the SIM (either beforehand or while the SIM is actually in use). Then, after an initialising operation (‘activating the purse’), the consumer can load the purse with electronic money and use it for making payments. The mobile phone SIM carries the purse application, while the mobile phone acts as the payment terminal that facilitates the communication between the e-purse application on the SIM and the SAM of the merchant.
C. E-purse linked to mobile phone number
In the third system there is no need to adapt the SIM. The consumer is offered a system that allows a remote purse application containing pre-paid value to be linked to the number of the mobile phone. The load procedure may take place by using scratch cards, direct debits, credit cards, voice response systems, etc. Through this procedure the pre-paid value bought is added to the sum total of the pre-paid value that is linked to the mobile phone number. The e-purse application and the related e-money are not stored on the mobile phone SIM itself but constitute a server-side wallet application. This wallet application, also called m-wallet, is maintained by the company that issued the pre-paid value. The mobile phone acts as the payment terminal that facilitates the use of the m-wallet on the remote computer server.
One of the issues on which the regulatory approaches differ is whether the concept of e-money requires the e-purse application and/or the electronic money itself to be in the logical possession of the holder (the functional approach) or to be in the physical possession of the holder (the physical possession approach).
The supervisor/regulator who applies the functional approach will notice that in all three systems the consumer is in full control of the purse application that holds the pre-paid value, even though this application may reside on distinct technical devices (IC chip, SIM, Internet server). All three systems perform the same function as they allow the pre-paid value to be spent at an organisation other than the issuer. Thus, even though the systems vary in their technical implementation, they are all e-money systems. A non-bank organisation operating such a system needs to apply for a license as an e-money institution, or it can use an exemption or waiver regime (if applicable).
The supervisor/regulator who favours the physical possession approach will argue that systems A and B are e-money systems, while system C is not. In system C the m-wallet is not on the mobile phone that is in the physical possession of the holder. Instead, the mobile phone number serves to identify the customer’s m-wallet on the remote computer server of the issuer. This server is not in the physical possession of the consumer. System C is therefore not an e-money system, contrary to systems A and B.
In a worst case scenario, the supervisor/regulator will view the activities in system C as a banking activity and impose bank license requirements. The result is that, of the three systems that are functionally identical, the organisation that has built system C (the most efficient and consumer-friendly system) is faced with the choice of whether:
- to apply for exemption under the regulations,
- to apply for or use small-scale waiver regimes (if applicable),
- to redesign its system into system A or B (making it more costly and less efficient),
- to operate under the regulatory bank regime imposed by the regulator/supervisor,
- to move its operations to a member state where regulators/supervisors have agreed on the functional approach(and use the European passport to re-enter the market).
This example hopefully illustrates what may happen when definitions and concepts of electronic money are interpreted differently by European regulators/supervisors. In view of the fast growth of the market for paid mobile services, there is an urgent need for regulators/supervisors to agree on a common understanding of electronic money. Such an understanding should not hamper efficient technological innovation and would establish harmonised supervision for all market players in and across the European member states.
In order to contribute to this common understanding, we present the following Annexes, which contain background information and analysis that underlie our conclusions:
Annex 1 - The concept and definition of electronic money,
Annex 2 - The players in the market for e-money,
Annex 3 - The local implementation of the e-money definition in the EMI-Directive,
Annex 4 - The current and future position of m-wallet providers.
We hope that this position paper clarifies the perspective of some of the e-money institutions in the Netherlands and that it may be of use in the discussions of market players, regulators and supervisors.
Annex 1: The concept and definition of electronic money Before we discuss the actual definitions of electronic money, we feel we need to discuss some underlying frameworks and the early e-money products. In section 1, we discuss the different concepts of banking. As our starting point in this discussion, we stress that the provision of payment instruments is a separate and profitable line of business. The discussion of early e-money products in section 2 serves to highlight that, while some of these products may have appeared to allow independent user-to-user transactions and transfers of e-money, they were in fact still based on a centralised technical design and operation. We elaborate on this issue, given the possibility that today’s discussions on e-money may still be influenced by the outward appearance of these early products than by their actual technical implementation.
In section 3 we will see that current definitions of e-money vary since banking supervisors, the Financial Action Task Force on Fraud (FATF) and payment overseers at central banks all have their own approach. The analysis shows that most definitions contain some technical bias, but the definition in the EMI directive does not. This means a good starting point exists for the conceptual discussion of e-money and its implementations in sections 4 and 5. We hope that the functional approach that we will introduce will be helpful in identifying different types of e-money products and the range of issuing/acquiring and clearing and settlement arrangements possible.
1.1 Banking: from provision of payments to taking deposits and back?
The word bank originally comes from the table (‘banca’) that was used by money exchangers to count, weigh and sort money of various quantities and qualities. As money mainly served as a payment instrument, definitions of banking in the 1800s stressed the provision of payment services as the prime banking activity. The establishment of central banks and the growth of national economies led to a change in the understanding of what constitutes a bank. The role of a bank shifted from that of providing payment instruments towards that of accumulating deposits/savings and of converting money titles (loans, savings, securities). As a result, even today the academic definition of a bank tends to focus on the function of taking deposits rather than the function of providing payments services/facilities.
In line with the more recent concept of banking, the current discussion on e-money focuses more on the role of e-money as a store of value than as a payment instrument. We think, however, that in order to conduct a proper discussion on the definition of e-money, it is necessary to be aware of the fact that money itself allows two separate types of businesses to exist. One is the payment services business (transporting value); the other is the deposit/loan/securities/savings business (storing and transforming value).
We recognize that historical conditions and industry dynamics may in some countries have led to a situation where deposit and payment business are combined or intertwined. Also, some regulatory frameworks may have their origin in the deposit-taking approach to banking and have been developed and modified on that basis. We also note that traditional banks sometimes claim that their payment business is unprofitable, thereby suggesting that the payments industry could not exist as a separate industry. Nevertheless, we wish to point out that both old and new technologies allow for efficient and profitable provision of payment services as such. We therefore recommend this as the explicit starting point for the discussion.
More specifically, the emergence of Payment Service Providers (that are no longer loss-making) demonstrates that new technology may indeed help to provide efficient payment solutions that contribute to the uptake and further development of electronic commerce. A similar role may be in store for the new e-money institutions.
1.2 Electronic money: the legacy of the early examples
Although chipcard-based systems were around for a while, the first generic electronic money systems came into existence in the early 1990s. The three main products were Danmont, Proton and Mondex. These products led central banks to formulate an opinion at the European level. A report written under the auspices of the European Monetary Institute (EMI, 1994) concluded, amongst other things, that the issuance of e-purses that could be used for a wide variety of purposes included the process of deposit-taking from the public and thus required a bank license.
Two years after the EMI report, an Internet-based form of e-money (e-cash by Digicash) was being developed and piloted on the web. One specific feature caught attention. E-cash was stored as digital coins on a personal computer and could be sent anywhere by personal computer (and by the related e-cash application). It appeared to be a system whereby individuals could keep on sending money back and forth over the Internet without central intervention. The e-cash product showed that not only card-based e-money existed, but also some kind of network-based e-cash systems. These new systems were labelled network money.
With the example of e-cash in mind, one of the identifying characteristics of network money systems in general was that a direct user-to-user cash payment function existed, without intervention by a central system operator. The technical reality of e-cash was different, however. It was technically impossible to immediately use received digital strings of e-cash for payment to any other e-cash accepting individual. In its technical implementation, every e-cash coin that was received was immediately sent to the central Mint, where it was checked, with new coins then being re-issued to the receiver of the payment. No coin was used more than once, and it was not possible to use the exact same coin (digital string) that had been received from another user.
Formally, both the ECB report on e-money (ECB, 1998) and the EMI directive (2000) acknowledge the need for a technology-neutral legal framework. However, some of the definitions and explanations that circulate leave the impression that some regulators/supervisors are still influenced by the external product characteristics of the early examples of e-money. This understanding may be technically incomplete and may impede understanding the technology-neutral definitions in a way that allows a wider range of technical choices to be made.
New electronic means of retail payment that are currently being tested or implemented in a number of markets include multi-purpose prepaid cards, sometimes called "electronic purses" or "stored-value cards", and prepaid or stored-value payment mechanisms for executing payments over open computer networks, such as the Internet. For the purposes of this report, these products are referred to as electronic money. A precise definition of electronic money is difficult to provide; indeed, a number of official bodies have described and categorised these products in different ways. When investigating some of the definitions/distinctions that are in use today, we still see no uniformity with respect to the definition of e-money. To illustrate this, we present the definitions of the Financial Action Task Force on Fraud (FATF), of the Payment Systems Policy Department of the European Central Bank and those in the EMI directive (1998/2000).
The FATF (2002, p. 19) states in its Review of the Forty Recommendations:
22.214.171.124. Electronic money (purses and cards)
67. The term electronic money designates a claim on the issuer of the money that is stored in an electronic medium and is accepted as payment by third parties other than the issuer. The electronic medium could be a smart card, in which case it is called an electronic purse. When the medium is a server run by the issuer and is accessible through the Internet from PCs running the appropriate software, it is referred to as a virtual purse. In the consultation paper on E-payments, the ECB (2002, p. 12) writes:
Several prepaid schemes have emerged in Europe for small-value e-payments. A distinction is made between three groups:
(i) “e-money schemes” which were originally developed to replace small cash payments in everyday life;
(iii) “prepaid cards” which were developed for anonymous and small-value payments over the internet. The first draft of the EMI directive (EC, 1998) stated:
'electronic money` shall mean monetary value which is
(i) stored electronically on an electronic device such as a chip card or a computer memory;
(ii) accepted as means of payment by undertakings other than the issuing institution;
(iii) generated in order to be put at the disposal of users to serve as an electronic surrogate for coins and banknotes; and
(iv) generated for the purpose of effecting electronic transfers of limited value payments. We conclude that a wide range of technical implementations of e-money has been and will be envisaged. The FATF focuses on the difference between card-based purses and server-based purses (wallets), both of which can be viewed as e-money. The ECB notes that e-mail payment mechanisms are a market reality as well and introduces the concept of personalised accounts. We have seen that the first draft of the EMI directive still hinted at specific technologies (card, computer memory); in its final version (2000) the definition became more neutral, however:
"electronic money" shall mean monetary value as represented by a claim on the issuer which is:
(i) stored on an electronic device;
(ii) issued on receipt of funds of an amount not less in value than the monetary value issued;
(iii) accepted as means of payment by undertakings other than the issuer. In addition, the EMI directive explicitly states in its considerations that the technology-neutral framework should assist electronic money in delivering its full potential benefits and should avoid hampering technological innovation. To public authorities the challenge to apply such a technology-neutral regulatory framework comes down to:
- differentiating between a functional and a technical view of ICT systems,
- appreciating the inherent flexibility of today’s technology.
To market players, the conceptual switch from functionality to technical implementation is one that is a regular business practice, as this is a formal stage during the development of ICT systems. Consequently, market players easily pick up the concept of technology-neutral implementation of definitions and quickly identify and cluster all kinds of technical implementations as e-money and their respective providers as e-money institutions.
As technological neutrality goes to the heart of the current debate, we will go on to describe in the following sections how one model of an electronic money system can be implemented in a variety of ways.
1.4 Functional model of an e-money system
The basis of our modelling is the general model presented in the BIS report on the Security of Electronic Money (BIS, 1996, p. 34). This model of e-money systems distinguishes three separate domains:
- the clearing and settlement domain, in which financial institutions, clearing houses and the central bank fulfil the interbank financial obligations resulting from electronic value transactions,
- the issuing/acquiring/operating domain, in which a structure is set up for issuing and acquiring electronic value as well as for interacting with the clearing and settlement domain; and
- the retail domain, in which the actual value transfers between users take place:
· loads : transfers of value from the issuer to users;
· payments : transfers of value between users;
· deposits : transfers of value from users to an issuer or an acquirer.
For our discussion we have adapted the general model, allowing us to focus on arrangements in the issuing/acquiring/operating domain and the retail domain (see Figure 1). We view the model as a functional model, as it identifies the main functions that need to be implemented for an e-money system to operate. The model does not imply any organisational or technical choices with respect to the functions identified. A definition of e-money, in line with this functional approach, would be:
electronic money is the representation of prepaid value on an electronic device.
Figure 1 – Functional model of an e-money system
In all domains of the model, choices have to be made as part of the practical design of the organisation and operation of an e-money system. The most important choice in the clearing and settlement domain is:
- should the e-money institution act as a separate payment institution in the clearing and settlement domain or use existing products/banks in order to provide the necessary interbank transactions?
- is the issuing and acquiring function centralised or does delegated issuing/acquiring exist?
In Figure 1 we have assumed that the e-money provider issues e-money to the consumers itself but allows for a flexible acquiring structure. In this structure, a number of organisations may resell the specific e-money product, provided that the merchants comply with the technical requirements of the e-money provider and sign an operational agreement with the e-money provider to this effect.
In a general sense, the model shows that in the issuing/acquiring/operating domain, e-money institutions need to implement the following functions:
I. Receiving pre-payment from consumers
II. Managing/administering the float (funds received)
IV. Receiving claims from merchants/consumers (redemption requests)
V. Paying the merchant/consumer
1.5 Implementation choices in the retail domain
Figure 1 shows that in the retail domain, the functions to be implemented are:
1. Loading the e-purse
2. Paying/receiving e-money
3. Redeeming e-money.
The choices that need to be made with respect to commercial characteristics of these functions are:
- the name, positioning and marketing of the e-money product,
- the limits applicable to payment/loading/redemption, which may differ per type of user,
- the legal qualification of the e-money payment (as a part of the contract terms),
- the use of personal data elements (name, e-mail address, phone number) for communication and transaction purposes.
In a technical sense, the challenge is to design an architecture that allows the e-money system to be used in a wide variety of ways. Detailed choices need to be made with respect to:
- the need/possibility of a shadow administration per e-purse,
- the tools/technologies that clients must use for communication and transactions (PC and Internet, mobile phone, interactive voice response, SMS),
- the technical representation of e-money (digital coins, balance, or a digital certificate in combination with a data record),
- the appearance of the e-money applications towards users (consumers may view it as an SMS system, while merchants may only see and use an Internet-based application),
- the specifications of the device(s) that carries the e-purse application and the balance of e-money (a PC, a mobile phone, a proximity token, an IC card).
The essential technical question is to design the E-purse application(s) architecture. The idea behind many of the older systems is that there is a decentral device (an IC chip or a PC) that carries a single application (the e-purse software) allowing the user to manage the content of the e-purse (load/redeem) and to effect payments. Experience has shown, however, that these decentralised systems are quite costly to operate. At some point during the lifetime of the product, improvements will need to be made, but these cannot be rolled out overnight for all devices. This results in the simultaneous operation of a number of releases of the same system. To avoid these migration costs and limit the backward compatibility problem, many industries (not just banks) are migrating towards ICT infrastructures in which applications are centrally hosted.
1.6 The trend to server-based architectures In order to illustrate the possible design choices, we describe three pre-paid electronic purse systems, all of which make use of electronic purse, chipcard and mobile phone technology. These are:
A- a dual slot e-purse system,
B- an e-purse residing in the SIM
C- an e-purse linked to the mobile phone number.
The first two of these systems are examples of a decentralised application architecture; C is a centralised architecture.
A. Dual slot e-purse system
The consumer uses the regular bank chipcard that has an electronic purse application. The consumer enters the chipcard into his/her mobile phone that is equipped with a second slot to read IC cards. The mobile phone operates here as the payment terminal that facilitates the communication between the purse application on the chipcard and the Security Application Module (SAM) of the merchant. This SAM validates and accepts the e-purse payments and may reside on a remote server of the merchant.
B. E-purse in Subscribe Identification Module (SIM)
In this set-up, banks and mobile operators have agreed not to use mobile phones with chip-card readers but to use a part of the consumer’s SIM to store the electronic purse application. As a result, the e-purse software must be loaded into the SIM (either beforehand or while the SIM is actually in use). Then, after an initialising operation (‘activating the purse’), the consumer can load the purse with electronic money and use it for payments. The SIM of the mobile phone carries the purse application, while the mobile phone acts as the payment terminal that facilitates the communication between the e-purse application on the SIM and the SAM of the merchant.
C. E-purse linked to mobile phone number
In this third system there is no need to adapt the SIM. The consumer is offered a system that allows a remote purse-application containing a pre-paid value to be linked to the number of the phone. The load procedure may be effected by using scratch cards, direct debits, credit cards, voice-response systems, etc. Through this procedure the pre-paid value bought is added to the sum total of pre-paid value that is linked to the mobile phone number. The e-purse application and the related e-money are not stored on the SIM of the mobile phone itself, but in fact constitute a server-side wallet application. This so-called m-wallet application is maintained by the company that issued the pre-paid value. This wallet application, also called m-wallet, is maintained by the company that issued the pre-paid value. The mobile phone acts as the payment terminal that facilitates the use of the m-wallet on the remote computer server.
Recent developments of electronic banking products show a move towards centralised application management and application service provision (ASP models). This trend can be observed in the electronic payments domain as well. All kinds of e-mail payment products as well as micro-payment products have a centralised technical architecture. In our view, the technical choice for centralised application management does not affect the functionality of the product. The functionality of an e-money application (or any other application, for that matter) on the hard drive of a local PC will not differ from that of the same application when run on the hard drive of a network server.
The conceptual challenge in the e-money discussion might be to incorporate the technical trend towards centralised application management. This facilitates a multi-channel approach towards the provision of services to consumers, allowing a wide range of technical instruments to interact with a central application. When applied to e-money, consumers are likely to be offered a number of ways to use their future electronic purse. The ‘device’ in the e-money discussion is therefore not necessarily the transaction device (as with electronic purses on IC cards), but it may also be the central computer server that centrally holds the e-purse application and e-money of a specific consumer.
1.7 The business perspective: is it e-banking or e-money?
The use of a functional approach may lead to the question as to the difference between an electronic banking application and an e-money application with a remote e-purse application that contains e-money. The answer to that question, in our view, does not lie in the identification of specific technical features of accounting or bookkeeping practices. Both e-banking systems and e-money systems will make use of similar security technology, apply accounting practices that allow individual integrity checks of records and balances and may technically be designed for multichannel usage (allowing remote applications). So what would determine whether a product constitutes e-money and its provider an e-money institution?
The answer may lie in the business function that the product and its provider intend to fulfil. If we consider that money may serve both as a payment mechanism and as a store of value, we can also identify whether features of a payment product and the business concept of its provider are those of a payment product provider or those of a regular credit institution.
A payment product provider will focus on the payment transaction between the consumer to a third party. Its business case will be based on providing customers a better payment product in terms of fee, speed, technical system applied, guarantee and ease of use. The monetary value in the payment system is limited to that of the transactions effected by its users. The business case will be based on transaction and contribution fees and not on the interest earned on the monetary value in the system. The legal qualifications of the payment service will specify the conditions for payment and its irreversibility/guarantee.
The credit institution will focus on the relationship between the customer and itself and specify the conditions of this relationship. Its business case will be based on providing the customer a range of financial products for storing value (savings, securities, loans), each with different characteristics in terms of price and risk. The interest margin on the value deposited will constitute a substantial part of the credit institution’s revenue. Furthermore, a part of the value that consumers have stored with the credit institution will at some future point in time be withdrawn by the consumers themselves (in the form of cash).
Interestingly enough, there is one particular feature that would essentially differ between the use of electronic value as a store of value or as a means of payment. That feature is, whether the value represented can or cannot be redeemed by the consumer. Any payment product provider will choose to make clear to its consumers that the value is meant to be spent with third parties and not to be redeemed, while a credit institution will always allow withdrawal of funds (as a basic service). From this perspective it is rather unfortunate that the EMI directive obliges all e-money issuers to redeem the pre-paid value upon request.
For a further discussion of this issue we refer to page 14, where we will see that the approach taken by the Financial Services Authority may be a good basis for a functional approach.
Annex 2: The players in the market for e-money, Having discussed the concept of electronic money and the issues regarding its implementation, we now look at the suppliers of e-money systems and products. We present an overview that is based on the Dutch market situation.