Digital Money and Monetary Control
Aleksander BERENTSEN <email@example.com>
This paper considers the implications of digital money for monetary control. For this purpose, it first investigates the potential of digital money to replace central bank currency. Considering its characteristics only, digital money could eventually become the dominant payment instrument for small value payments. The main obstacles to its success, however, are network externalities. These externalities are discussed in a simple game-theoretic model in which buyers and sellers decide whether to use digital money or central bank currency for payments. The paper then studies the implications of a complete replacement of central bank currency. Here, the main problems for central banks are the loss of seigniorage income and the nontrivial reduction of total liabilities and assets, which adversely affect monetary control. It ends with a discussion of the measures that central banks can take to prevent these developments.
The possibility that consumers could start paying for all retail purchases with digital money,1 once it becomes established, instead of with paper money, has begun to alarm central bankers, the media, and scholars and has recently initiated a number of studies. The main examples are two papers published by the Bank for International Settlements (BIS 1996a and BIS 1996b). BIS (1996b) considers the implications of the development of digital money for central banks in such areas as monetary policy, supervision and regulation of financial intermediaries, and supervision of the payment system.
BIS (1996a) focuses on the security of digital money products, which is the major concern of central bankers today. A security breach -- counterfeiting -- of a widely used digital money product could rock the financial system. In addition, there is concern that digital money could facilitate money laundering, fraud, and tax evasion; illegal activities could be facilitated by widespread use of digital money because, in contrast to debit card and credit card transactions, some forms of digital money allow users to remain anonymous. For banking supervisors, the main issues are whether institutions, other than banks, should be allowed to issue digital money, and whether traditional regulations, such as reserve requirements and capital regulations, should be extended to all issuers of digital money. The European Union, for example, plans to limit the issuance of multipurpose smart cards to registered banks only (EMI 1994).
Digital money proposals are based on either the smart card or computer network money. The smart card (also known as the electronic purse) is a plastic card with an embedded microprocessor that can be loaded with a monetary value. The card's value is reduced with each purchase. Reloadable, the smart card needs no online authorization for value transfer and can be used for many different purposes. In contrast, the single-purpose, prepaid cards (e.g. telephone cards) widely used in Europe are neither reloadable nor can they be used for multiple purposes. The main difference between the smart card and debit card is that the use of a smart card initiates a transfer of a medium of exchange, i.e., digital money, that is stored on the smart card. In contrast, the use of a debit card initiates a transfer of a medium of exchanges, i.e., demand deposits, which is stored on the transaction account of the payer. Thus, for a debit card payment an online connection is needed. Basically, network money is software that allows the transfer of value on computer networks, particularly on the Internet.
Like a demand deposit, digital money is inside money, whereas central bank currency is outside money. Inside money gives the bearer a legal claim against its issuer; outside money entails no such claim (Hellwig 1985). Like a travelers check, which is also inside money, a digital money balance is a floating claim on a private bank or other financial institution that is not linked to any particular account (White 1996). The incentive for an institution to issue digital money is the interest-free or low-interest debt financing that its outstanding balance provides.
This paper examines how widespread use of digital money would affect monetary control. It considers digital money's potential to replace paper currency because its principal threat to monetary control stems from its potential to replace central bank currency, which in most countries is by far the largest component of central bank liabilities. We also evaluate possible central bank responses to the development and spread of digital money.
Digital money products, based on smart cards, are designed to facilitate small-value payments in face-to-face retail transactions. The average transaction is expected to be less than $20, a sum for which use of credit and debit cards is inconvenient and too costly (CBO 1996). It is therefore expected that digital money based on smart cards will reduce use of central bank notes and coins and, to a much lesser extent, use of debit and credit cards for face-to-face payments. Software-based digital money products are more likely to reduce use of checks, debit cards, and credit cards for non-face-to-face payments, i.e., for online payments.
Many different types of digital money with varying characteristics are currently being developed. In principal, however, digital money can be designed to share all characteristics of central bank currency, and can be seen, therefore, as a very close substitute for bank notes and coins. Table 1 lists the main characteristics of currency, digital money, checks, and debit cards. The main differences between digital money and currency are:
Replacement of currency could benefit issuers, consumers, and merchants. Issuers would benefit from the interest-free debt financing provided by digital money balances. Today, almost all proposed digital money schemes are developed by private institutions. The issuer, however, could also be a central bank. Consumers would benefit from the convenience. For merchants, accepting digital money would reduce costs if bits and bytes replaced physical coins and notes.
The estimated annual costs of handling central bank currency by U.S. retailers and banks are $60 billion; this total includes costs of processing and accounting of money, storage, transport, and security (Hayes et al. 1996). Hayes et al. (1996) also suggests that the cost of an electronic payment ranges between one-third and one-half of a paper check or paper giro payment. Beside cost savings, Hayes et al. (1996) suggest additional advantages of using digital money including an expected reduction in some forms of fraud, greater safety and security, and a potential for value-added services. An indication of the low cost per transaction: several money schemes under development would allow (online) payment for services (e.g. downloading a newspaper article) that cost less than one cent (micropayments).
Table 1 compares the characteristics of digital money, currency, checks, and debit cards. In many respects digital money has the same characteristics as currency. However, it is the only payment instrument that can provide payment finality in non-face-to-face transactions. This property and the low costs per transaction make digital money the ideal candidate for payments across public computer networks.
A property not mentioned in Table 1 is robustness of a payment instrument against fraud, technical problems, or changes such as the year 2000 problem and the introduction of the Euro. In this respect the rating would be less favorable for digital money. An anonymous referee suggested a rating of No, Yes, Yes, and yes. The main reason for the poor rating of digital money with respect to this property is that after an unexpected change the software on all smart cards and computers would have to be exchanged. The decentralizes nature of digital money makes such changes very costly. In a more centralized payment system, such as the credit card business, the costs of such changes are expected to be less.
While digital money products based on smart cards would mainly reduce demand for currency, software-based digital money products could also affect demand for demand deposits due to reduced transaction costs and learning spillovers. Software-based digital money could facilitate and reduce the cost of transferring value among different types of accounts, banks, and countries.
The second reason -- learning spillovers -- refers to the notion that using software-based digital money will improve the skills of those using personal finance software and telecommunication technologies to optimize their financial planning. Personal finance computer software tracks income and outgo and helps in creating budgets, managing investments, and filling out tax forms. It also can access electronic banking and electronic bill payment services.
With these software products, small businesses and households will be able to monitor money holdings more frequently and to automate basic tasks.2 One day, for example, software agents could possibly automatically invest short-term excess liquidity overnight for ordinary households. It is difficult to predict the quantitative effects of these developments on the demand for transaction deposits. However, lower transaction costs and more frequent adjustments of money holdings suggest a lower demand for demand deposits.
This comparison of digital money and central bank currency suggests that digital money is a serious competitor to central bank currency. It is possible, therefore, that digital money could replace the entire stock of central bank currency. The main obstacles to its success, however, are network externalities. The benefit of using any money product depends on its popularity. If few merchants accept digital money, households using digital money will find few benefits. Similarly, if few consumers use digital money, merchants will have little incentive to accept digital cash. These network externalities suggest that even though use of digital money would benefit firms and households, paper currency will not be replaced easily. Issuers of digital money schemes will have to convince consumers and merchants that their product will succeed in the market against other digital money schemes as well as those other popular and traditional media for small value payments, checks and cash.
In the following we consider a simple game theoretic model that captures these network externalities.3 The starting point, i.e., the status quo, is the existence of a generally accepted medium of exchange used to pay for all purchases in the economy. At some point, a private entity launches a new type of money with favorable properties for shop owners and consumers. Use of the new money requires an initial investment by both firms and consumers. Each agent has to decide whether he or she wants to use the new money. The rate of return on the investment depends on how many agents decide to invest. Thus, like a telephone network, the incentive to participate depends on the number of users. If only a few merchants accept digital money, why would a consumer acquire a smart card or network money? At the same time, if only a few consumers are using smart cards, why would a merchant invest in equipment to process digital money?
The model is simple: There is one consumer and one shop owner and one transaction between the two.4 Both parties have to choose whether to use the old or the new money to pay for the transaction. Old money is denoted by M1 and is called cash. Use of cash costs a1 for both consumer and shop owner. The new medium, named digital cash, is denoted by M2. Use of digital cash requires from both parties an initial investment expenditure f. Use of digital cash costs a2 for both consumer and shop owner. Digital cash has superior characteristics for both merchant and consumer; in particular, total costs for both parties are smaller when they use digital cash, i.e., a1 > f + a2.
Of course, profitability of investment depends on whether both parties use digital cash. The probability that the shop owner invests in digital cash is denoted by and the probability that the consumer invests in it is denoted by . When P denotes the revenue of the transaction, the profit function for the shop owner is:
The shop owner chooses to maximize (1). The best response correspondence is given by the following rule:
If , the shop owner does not invest;
If , the shop owner invests in digital cash; and
If , the shop owner is indifferent.
Note: The larger the investment, f, and the larger the transaction cost of digital cash, a2, the less likely that the shop owner will invest in digital cash. Also, the larger the transaction cost of cash, a1, and the larger the probability that the consumer will use digital cash, , the more likely that the shop owner will invest in digital cash, the new medium of exchange. Next, consider the consumer who has to decide whether to make the investment. The consumer's expenditure function is:
The consumer chooses to minimize expenditure e. The best response correspondence is given by the following rule:
If , the consumer does not invest;
If , the consumer invests in digital cash; and
If , the consumer is indifferent.
Figure 1: Best response correspondences of shop owner (dashed curve) and consumer (solid curve) who have to decide whether to use the new media of exchange. The best response of one party depends on the planned action of the other party.
The best response correspondences of the consumer and of the shop owner are shown in Figure 1. There are three (Nash) equilibria in this game. In the first equilibrium, A, digital cash is not used, even though it would be less costly for both parties. In the second equilibrium, B, cash and digital cash are used simultaneously and in the last equilibrium, C, only digital cash is used.
Note further that, even with zero investment cost equilibria, A prevails. If the consumer does not use digital cash, a best response for the shop is not to accept digital cash. And, given that the shop does not accept cash, a consumer's best response is not to use cash. We have kept the model simple, but even in a more realistic setup the basic problem of strategic interaction remains: The behavior of one party depends upon the expected action taken by the other party.
The model suggests that even when replacement of central bank currency would benefit firms and households, cash wouldn't be replaced as easily. Issuers of one kind of digital money would have to convince consumers and shop owners that their investment would have a positive rate of return; basically, this means that their particular technology would succeed in the market against other digital money schemes and traditional media. It also suggests that convincing agents to use digital money will make implementing digital money schemes costly.
Monetary control is based on the ability of central banks to determine the conditions that equilibrate demand and supply in the market for bank reserves. In this market central banks are monopolistic suppliers of reserve assets and they can also directly affect demand, for instance, by setting reserve requirements and by shaping and operating key interbank settlement systems (Borio 1997). Central banks can either decide to control the total quantity of reserves or the price at which they are traded among banks. In most countries central banks aim at stabilizing the short-term interest rate at which banks trade these reserves. In the following we study the impact of digital money on the demand for bank reserves (deposits at the central bank) and discuss the implications for monetary control.
Banks hold reserves for two reasons. (1) In many countries they are required to hold a percentage of certain types of deposits as reserves. The percentage and the types of deposits that require holding reserves differ from country to country. (2) Banks hold reserves for settlement purposes to cushion costly daytime and overnight overdrafts. The types of liquid assets that are counted as reserves differ from country to country. Most reserves, however, are held as book entries at the central bank and as vault cash, i.e., central bank currency holdings of banks.
In the market for reserves, banks trade reserves to meet reserve requirements and to adjust settlement balances. The price for these reserves is the primary interest rate with which central banks influence through their monetary instruments. A permanent change in the price for reserves influences other money and credit rates and, eventually, the real sector of the economy and the price level.
Demand for reserves depends crucially on the institutional arrangements prevalent in a country and these institutional arrangements differ from country to country. They pertain to the existence of (binding) reserve requirements, the nature of payment and settlement procedures, the types of eligible reserve assets, and the conditions of central bank assistance (standing facilities). For an excellent, detailed description of institutional arrangements in several industrial countries consider Borio (1997).
The banking systems' demand for reserves to meet reserve requirements depends on the public's demand for reservable deposits and reserve ratios. Demand for reserves would be reduced if digital money were to become a substitute for reservable deposits. Section 1 suggested that software-based digital money could reduce demand for transaction deposits. The Bank of International Settlements (1996b) suggests, however, that substitution of reservable deposits would be small: "It is conceivable that a very extensive substitution [of reservable deposits] could complicate the operating procedures used by central banks to set money market interest rates. However, since e-money is expected to substitute mostly for cash rather than deposits, it is highly unlikely that operating techniques will need to be adjusted significantly."
During the 1990s, many industrial countries have radically reduced their reserve requirements. For example, since 1990, reserve requirements have been reduced in all of the major (G7) industrial countries (Bisignano 1996). Today, Belgium and Sweden have no reserve requirements in place. Digital money could reduce demand for reserves indirectly by adding to this trend, particularly those software-based products that facilitate and reduce the cost of transferring value among different types of accounts, banks (and nonbanks), and countries.
If bank customers were to use digital money extensively, the pressure on central banks to reduce reserve ratios and the number of types of reservable liabilities would increase. This would be particularly true if foreign intermediaries were to increasingly attract (transaction) deposits of domestic residents across public computer networks; central banks would be pressed then to lower reserve ratios to help domestic banks compete for domestic (and foreign) deposits.Reserve requirements are basically a tax on financial intermediation; banks subject to reserve requirements are at a competitive disadvantage compared with nonbank financial intermediaries offering close financial substitutes. A study by the Bank of Japan (1995) on the recent reform of reserve requirements in major industrial countries suggests that the main reason for reducing reserve requirements has been to lower the "burden" ("distortion", "inequality") on depository institutions (Bisignano 1996).
Without reserve requirements or where reserve requirements are nonbinding, demand for bank reserves is essentially a demand for settlement balances. Demand for reserves would be affected if digital money were to change the need for settlement balances. In the absence of reserve requirements, a bank would still want to hold reserves to meet unforeseen liquidity withdrawals by its customers. However, the holding of reserves is costly and banks have to choose the optimal amount of reserves to hold.
This basic liquidity management problem is presented in Baltensperger (1980).4 During a given period, a bank faces a stochastic outflow of deposits, X, with density function, f(X). At the beginning of the period the bank has reserves, R. If at the end of the period , the bank has a reserve deficiency and has to bear the cost ; p measures the cost per unit reserve deficiency. Holding reserves is costly, too, because reserves could alternatively be invested and earn a rate of return of r.
In the absence of reserve requirements, the bank chooses its reserve level so that the marginal cost of reserve holdings, r, equal the marginal return of holding additional reserves, i.e.,
Digital money would affect the demand for reserves if it were to change the cost of reserves, r, the unit cost of reserve deficiency, p, or the probability that a reserve deficiency occurs . There is no reason to believe that digital money per se would affect either of these parameters. The rate of return on loans, r, is determined by demand and supply in the market for loans and should not be affected by the emergence of digital money. Equally, the unit cost of reserve deficiency, p, is determined by the conditions of central bank assistance and the cost of selling illiquid assets at short notice, and should as well not be affected by digital money. Finally, digital money per se should not affect the probability that a reserve deficiency occurs.
Thus, the liquidity management approach suggests that digital money would not reduce demand for settlement balances importantly. This implies that, even with zero reserve requirements or even when digital money extensively substitutes for reservable deposits, the settlement function of central banks would continue to guarantee an ongoing demand for reserves. Reserves may not be used as a medium in which to store value overnight and longer, but it still would be required as a vehicle for transferring value during a day (Jordan-Stevens 1996). Havrilesky (1987) suggests that this demand would be sufficient for the monetary base to remain a viable policy instrument and that open-market operations would work as they do today.
Initially, substitution of central bank currency would increase cash holdings of banks since customers would return excess cash. Banks would observe that their cash holdings exceeded the optimal amount and they would return cash to the central bank, thereby increasing their reserves on the books at the central bank. Consequently, substitution of central bank currency would increase the total supply of reserves. Substitution of central bank currency, therefore, is equivalent to an expansionary open market operation that provides additional reserves to the banking system (Berentsen 1997).
Central banks could be forced to step in and absorb these reserves by selling central bank assets. Because currency is by far the largest liability of central banks, an extensive substitution could reduce the monetary base to the extent that it would adversely affect monetary control. This concern has been raised by BIS (1996b):
"Since cash is a large or the largest component of central bank liabilities in many countries, a very extensive spread of e-money could shrink central bank balance sheets significantly. [...], special circumstances could arise in which the central bank might not be able to implement reserve-absorbing operations on a large enough scale (for example, to sterilize the effects of large purchases in the foreign exchange markets) because it lacked sufficient assets on its balance sheet".
The special circumstances mentioned in the BIS report refer mainly to times when exchange rate commitments of central banks are challenged by market forces. To keep exchange rates in line, central banks intervene in the foreign exchange markets by either buying or selling foreign currency in exchange for domestic. These exchange rate operations affect the system's liquidity. In fact, the net creation of liquidity through foreign channels can be huge, amounting in some cases to large fractions of the outstanding stock of policy instruments (Borio 1997).
When exchange rates come under upward pressure, the main problem is to absorb excess liquidity created by the selling domestic currency in exchange for foreign. The main risk is that the central bank may not have enough assets to sell to withdraw the liquidity. When exchange rates come under downward pressure, central bank purchases of domestic currency for foreign absorb liquidity. In this case, the risk is that central banks may not succeed in injecting sufficient funds to meet the minimum settlement balance needs of banks, thereby effectively losing control over short-term rates and disturbing the settlement process (Borio 1997).
If all currency were to be replaced, the potential reduction of total liabilities would be nontrivial, as can be seen in the six industrial countries presented in Table 2. The table's first column lists total liabilities, the second column total currency, and the third column the currency-to-total-asset ratio. In Germany, for example, a replacement of the Deutsche Mark would shrink total liabilities of the Deutsche Bundesbank by 70 percent. In the United States, a complete substitution of central bank notes would reduce total liabilities and assets by 87 percent.
Besides affecting monetary control, replacement of central bank currency would reduce seigniorage income of central banks considerably. In many countries, central bank independence is based on the ability to generate more than sufficient seigniorage income to pay for operations. There is some concern that an extensive substitution of central bank currency could reduce seigniorage revenue to the extent that central banks would have to turn to other income sources such as government subsidies. However, Boeschoten-Heblink (1996), who studies the potential seigniorage loss of central banks in the G-10 countries, suggests that the share of seigniorage income used by central banks to pay for their expenses is small. This implies that even an extensive substitution of central bank currency would not force central banks to rely on other income sources.
Nevertheless, it is likely that central banks will take measures to prevent the loss of seigniorage income and to avoid the adverse effects on monetary control of a replacement of central bank notes and coins. Central banks could take the following measures:
Legal restrictions to prevent proliferation of digital money will be difficult to justify, especially in light of efforts to deregulate and improve the efficiency of the financial sector. Moreover, measures that prevent development of digital money products will result in a competitive disadvantage; nations that do develop these products will then be able to take the lead in a crucial technological sector. In addition, because digital money can easily cross international borders, it will be difficult to prevent the use of foreign digital money products that could eventually emerge as a new medium of exchange in the home country.
Central banks could provide digital money in the same way as they provide paper currency right now. The Bank of Finland, for example, is developing a cash-card system through its corporate subsidiary, Avant Finland Ltd. (Bernkopf 1996). Most central banks, however, have remained passive in this respect. There is concern that central banks issuing digital money products could limit competition and reduce incentives in the private sector to innovate more digital money products.
Central banks could require reserves on digital money balances. Reserve requirements, which are a tax on digital money, could reduce the private sector's incentive to issue digital money and could hold back or even prevent the private sector's incentive to invest in the development of digital money products.
In addressing digital money's potential to replace central bank currency, this paper suggests that digital money poses a credible threat to paper currency because it is specifically designed to make small value payments in retail transactions and because it could reduce transaction costs for consumers and businesses.
The main obstacle to its success, beside legal restrictions, are network externalities, i.e., the benefit of using a particular digital money product depends on the number of its users. If few merchants accept digital money, the benefits to households to use it are low. Similarly, if few consumers use digital money, a merchant has little incentive to accept it. These network externalities suggest that even though use of digital money could benefit firms and households, paper currency would not be replaced easily. Issuers of digital money will have to convince consumers and shop owners that their product will succeed in the market against competing digital money schemes and traditional media.
In examining the impact of digital money on demand for bank reserves, this paper suggests that digital money could reduce, to some extent, demand for bank reserves to meet reserve requirements. Demand for settlement balances is likely to be unaffected because digital money does neither affect the need for settlement balances nor does it affect the opportunity cost of holding reserves. This implies that both the settlement function of central banks and their reserve requirements would guarantee an ongoing demand for reserves.
The main effect of digital money would be on the supply side, i.e., on monetary control. In most countries, currency is by far the largest component of the total liabilities of central banks. Replacement of central bank currency would extensively shrink total liabilities (and consequently total assets) and limit the central banks' ability to control the economy. In particular, when market forces challenge exchange rate commitments, central banks may not be able to successfully sterilize liquidity generated by foreign market intervention.
Section 3 considered three possible reactions of central banks: They could legally limit the proliferation of digital money products, they could issue digital money themselves, or they could demand high reserve requirements on digital money balances. The drawback of these measures is that they would reduce the private sector's incentives to invest in developing digital money products. However, because of expected losses of seigniorage income and predicted adverse effects on monetary control, it is likely that central banks will resort to one or several of these measures. Initially, it is most likely that they will remain passive and observe whether digital money transactions will gain a nontrivial share among all payments. If so, they could try to impose legal restrictions on the use of this new payment instrument and impose reserve requirements on digital money balances to hold back its further development. In the long run, however, some central banks might also start issuing digital money themselves.
* I would like to thank an anonymous referee, Yvan Lengwiler, Daniel Heller, Marianne Bürgi, Michel Peytrignet, Daniel Rubinfeld, and Kevin Siegel for their helpful suggestions and the Swiss National Science Foundation for its financial support. I'm particularly indebted to Heidi Seney whose comments have improved this paper considerably. This paper draws on my article "Monetary Policy Implications of Digital Money", which appeared in Kyklos (International Review of Social Science) 51, Fasc.1, 89-117.
1. Literature on digital money has created many other expressions such as e-money, electronic money, network money, digital currency, electronic currency, digital cash, electronic cash, e-cash, etc. For an extended definition and characterization of digital money see BIS (1996a, 1996b) Stuber (1996), or CBO (1996).
2. Software agents are intelligent programs that can simplify the processing, monitoring, and control of digital transactions by automating many activities.
3. Consider also Dowd and Greenaway (1993) for a model of currency competition and network externalities.
4. The model can be extended in several directions. The set up could be made dynamic such that agents would repeadedly use a given payment instrment. On could also introduce several competing payment instruments which, for example, would differ in their cost structure.
5. Many authors have modeled in different variations the liquidity management problem of banks (see Baltensperger 1980). The discussion that follows is based on his exposition.
Baltensperger, E., 1980: Alternative Approaches to the Theory of the Banking Firm, Journal of Monetary Economics 6, 1-37.
Bank of Japan, 1995: Reserve Requirement Systems and Their Recent Reforms in Major Industrialized Countries: A Comparative Perspective, Quarterly Bulletin, May.
Berentsen, A., 1997a: Digital Money, Liquidity, and Monetary Policy, First Monday - The Peer-Reviewed Journal on the Internet, http://www.firstmonday.dk/issues/issue2_7/berentsen/index.html, Munksgaard International Publishers, Copenhagen.
Berentsen, A., 1998: Monetary Policy Implications of Digital Money. Kyklos (International Review of Social Science) 51, Fasc.1, 89-117.
Bernkopf, M., 1996: Electronic Cash and Monetary Policy, First Monday (Peer Reviewed Journal on the Internet), http://www.firstmonday.dk.
BIS, 1996a: Security of Digital Money, Bank of International Settlements, Basel.
BIS, 1996b: Implications for Central Banks of the Development of Digital Money, Bank of International Settlements, Basel.
Bisignano, J., 1996: Varieties of Monetary Policy Operating Procedures: Balancing Monetary Objectives with Market Efficiency, Working Papers No. 35, Bank for International Settlements, Basel.
Boeschoten, W.C. and G.E. Heblink, 1996: Digital Money, Currency Demand, Seigniorage Loss in the G-10 Countries, De Nederlandsche Bank Staff Report, May 1996.
Board of Governors (BoG) of the Federal Reserve System, 1994: The Federal Reserve System: Purpose & Functions, Washington, D.C.
Borio, C., 1997: Monetary Policy Operating Procedures in Industrial Countries, Working Papers No. 40, Bank for International Settlements, Basel.
CBO, 1996: Emerging Digital Methods for Making Retail Payments, the Congress of the United States Budget Office, Washington D.C.
Dowd, K. and D. Greenaway, 1993: Currency Competition, Network Externalities and Switching Costs: Towards an Alternative View of Optimum Currency Areas, The Economic Journal 103, 1180-1189.
EMI, 1994: Report of the Council of the European Monetary Institutions on Prepaid Cards.
Havrilesky, T., 1987: Monetary Modeling in a World of Financial Innovations, in E.H. Solomon (Ed.), Digital Funds Transfer and Payments: The Public Policy Issues, Kluwer, Boston.
Hayes et al., 1996: An Introduction to Digital Money Issues, prepared for the United States Department of the Treasury Conference "Towards Digital Money and Banking: The Role of Governments", Washington D.C.
Hellwig, M. F., 1985: What do we know about currency competition? Zeitschrift für Wirtschafts- und Sozialwissenschaften 105, 565-88.
Jordan, J.L. and E.J. Stevens, 1996: Money in the 21st Century, Cato Institute's 14th Annual Monetary Conference, http://www.cato.org/moneyconf/.Moody's International Manual, 1996: Moody's Investor Service, Inc., New York.
Stuber, G., 1996: The Digital Purse: An Overview over Recent Developments and Policy Issues, Bank of Canada.
White, L. H., 1996: The Technology
Revolution and Monetary Evolution, in The Future of Money
in the Information Age, Cato Institute's 14th
Annual Monetary Conference, http://www.cato.org/moneyconf/.
Tel. +41 31 631 80 80