University of Southern California Dissertations and Theses

Digital money

(USC Thesis Other)

Digital money

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content INFORMATION TO USERS This manuscript has been reproduced from the microfilm master. UMI films the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter face, while others may be from any type o f computer printer. The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adversely afreet reproduction. In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion. Oversize materials (e.g., maps, drawings, charts) are reproduced by sectioning the original, beginning at the upper left-hand comer and continuing from left to right in equal sections with small overlaps. Each original is also photographed in one exposure and is included in reduced form at the back o f the book. Photographs included in the original manuscript have been reproduced xerographically in this copy. Higher quality 6” x 9” black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact UMI directly to order. UMI A Bell & Howell Information Company 300 North Zed) Road, Ann Aibor MI 48106-1346 USA 313/761-4700 800/521-0600 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Digital Money Copyright 1998 by Georg Heinrich Strasser A Thesis Presented to the FACULTY OF THE GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment o f the Requirements for the Degree MASTER OF ARTS (Economics) August 1998 Georg Heinrich Strasser Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. UMI Num ber: 1 3 9 3 1 8 8 UMI Microform 1393188 Copyright 1999, by UMI Company. All rights reserved. This microform edition is protected against unauthorized copying under Title 17, United States Code. UMI 300 North Zeeb Road Ann Arbor, MI 48103 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. U N IV ER SITY O F S O U T H E R N C A L IF O R N IA T H E G R A D U A T E S C H O O L U N IV E R S IT Y P A R K L O S A N G E L E S . C A L IF O R N IA 90007 This thesis, written by Georg Heinrich Strasser under the direction o f hJL$ Thesis Committee, and approved by all its members, has been pre sented to and accepted by the Dean of The Graduate School, in partial fulfillm ent of the requirements for the degree of Master of Arts T in i^ A u g u s t 1 8 , 1 9 9 8 THESIS COMMITTEE Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. To my parents Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table of Contents 1 In God We Trust - All Others Pay Cash............................................................................. 1 1.1 The functions of money...................................................................................................1 1.2 Money on the balance sheet...........................................................................................2 1.3 Why has money value?.................................................................................................. 3 1.4 The wide variety o f money.............................................................................................4 1.5 Definitional problems..................................................................................................... 5 1.6 Private money................................................................................................................... 5 2 Digital money......................................................................................................................... 7 2.1 A primer on encryption.................................................................................................. 7 2.2 Electronic value..............................................................................................................11 2.3 Card-based digital money............................................................................................. 13 2.4 Network-based digital money..................................................................................... 15 2.5 Definition of digital money..........................................................................................18 2.6 Example of a digital cash transaction......................................................................... 20 3 Necessity of digital money................................................................................................. 26 3.1 Motivation for digital money...................................................................................... 26 3.1.1 Anonymity............................................................................................................... 26 3.1.2 Security.....................................................................................................................27 3.1.3 Lower cost o f money..............................................................................................28 3.1.4 Convenience............................................................................................................29 3.1.5 More efficiency - less friction.............................................................................. 30 3.2 Factors determining its success................................................................................... 33 3.2.1 Government payments........................................................................................... 33 3.2.2 Trust......................................................................................................................... 34 3.2.3 Convenience............................................................................................................35 3.3 The demand for digital money.....................................................................................37 3.3.1 Market environment...............................................................................................37 3.3.2 Market potential..................................................................................................... 38 3.3.3 A new choice...........................................................................................................41 3.3.4 Precautionary and speculative demand............................................................... 44 3.3.5 Transaction demand...............................................................................................46 3.4 The supply of digital money......................................................................................... 51 3.4.1 Incentives..................................................................................................................51 3.4.2 Costs......................................................................................................................... 53 3.4.3 Comparison............................................................................................................. 54 3.5 Current development status..........................................................................................54 3.5.1 Specific projects..................................................................................................... 54 3.5.2 General trends..........................................................................................................56 iii Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 4 Economic qualities of digital money.................................................................................59 4.1 Security and fairness..................................................................................................... 59 4.1.1 Counterfeiting, fraud and theft.............................................................................. 59 4.1.2 Contractual/ legal relationships............................................................................. 63 4.1.3 Consumer protection and disclosure.................................................................... 64 4.1.4 Access........................................................................................................................66 4.2 (Un-)traceability............................................................................................................. 67 4.2.1 Accounted and unaccounted systems................................................................... 67 4.2.2 Privacy.......................................................................................................................68 4.2.3 Money laundering....................................................................................................70 4.2.4 Taxation.................................................................................................................... 71 4.3 Competition.....................................................................................................................73 4.3.1 Low entry barriers...................................................................................................73 4.3.2 Ease of switching bank and currency................................................................... 74 4.3.3 Money rating............................................................................................................75 4.3.4 Information implications........................................................................................76 4.4 Transnationality.............................................................................................................. 77 4.4.1 Increase in cross-border payments....................................................................... 77 4.4.2 Competing national currencies..............................................................................78 4.4.3 Supranational legal systems.................................................................................. 80 4.4.4 Law enforcement.....................................................................................................81 4.5 Instability......................................................................................................................... 82 4.5.1 Speed.........................................................................................................................82 4.5.2 Operational risk....................................................................................................... 82 4.5.3 Squeezes................................................................................................................... 83 4.5.4 Noise and pre-selection..........................................................................................83 4.5.5 Mass speculation.....................................................................................................84 4.5.6 Confidence issues....................................................................................................88 4.5.7 Systemic risk............................................................................................................89 4.5.8 Real time settlement............................................................................................... 91 5 Digital money in macroeconomic models.........................................................................93 5.1 The Cagan model............................................................................................................93 5.1.1 Model setup..............................................................................................................93 5.1.2 Inflation.................................................................................................................... 94 5.1.3 Seigniorage...............................................................................................................95 5.2 A Sidrauski model..........................................................................................................95 5.2.1 Model setup..............................................................................................................95 5.2.2 Money demand........................................................................................................ 97 5.2.3 Stability.................................................................................................................... 98 5.2.4 Credibility.................................................................................................................99 5.3 Other models.................................................................................................................100 iv Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 6 Impact on central bank policy...........................................................................................102 6.1 Money supply............................................................................................................... 102 6.1.1 The monetary aggregates...................................................................................... 102 6.1.2 The behavior of the money aggregates............................................................... 107 6.1.3 Velocity of money.................................................................................................119 6.2 Seigniorage....................................................................................................................121 6.3 Control............................................................................................................................123 6.3.1 Shrinking central bank balance sheet................................................................. 123 6.3.2 More power for financial markets.......................................................................124 6.3.3 International issues................................................................................................125 6.3.4 Disintermeditation.................................................................................................125 6.3.5 Time lags................................................................................................................ 126 6.4 Confidence.....................................................................................................................127 6.5 Banking supervision.................................................................................................... 128 7 Policy options...................................................................................................................... 130 7.1 Freely competitive digital monies............................................................................130 7.1.1 Laissez-faire......................................................................................................... 130 7.1.2 From no intervention to free banking................................................................130 7.1.3 Private suppliers of digital money..................................................................... 136 7.1.4 A free private money market.............................................................................. 138 7.1.5 A central bank without power?...........................................................................147 7.2 Supervision of competing monies............................................................................. 148 7.2.1 Necessity of regulation.........................................................................................148 7.2.2 Technical restrictions........................................................................................... 150 7.2.3 Banking regulation and supervision................................................................... 151 7.2.4 Clearing.................................................................................................................. 154 7.2.5 Central bank lending............................................................................................ 154 7.2.6 Loss of control....................................................................................................... 155 7.2.7 Central bank monetary policy?............................................................................156 7.2.8 International coordination....................................................................................158 7.3 Centralized monetary systems....................................................................................158 7.3.1 Prohibition?............................................................................................................158 7.3.2 Central bank as issuer in competition................................................................159 7.3.3 Central bank as issuer in competition with distinct base money....................160 7.3.4 Central bank as exclusive issuer..........................................................................164 7.4 Critique.......................................................................................................................... 165 8 Conclusion............................................................................................................................166 9 Bibliography........................................................................................................................167 v Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. List of Tables Table 3.1: Cost per transaction by channel.............................................................................. 28 Table 3.2: Preferred forms of consumer payment in the USA.............................................. 38 Table 3.3: Payments in the USA in 1995..................................................................................39 Table 3.4: The precautionary demand for m oney...................................................................45 Table 6.1: Impact of digital money on an expanded M l...................................................... 112 Table 6.2: M l elasticity for various countries........................................................................116 Table 6.3: Currency in circulation in various countries....................................................... 117 Table 6.4: Seigniorage loss.......................................................................................................123 List of Figures Figure 2.1: Institutional arrangements for digital money.......................................................20 Figure 3.1: Digital money versus cash......................................................................................30 Figure 3.2: Economic effects of network payments................................................................32 Figure 3.3: Concurring advantages........................................................................................... 42 Figure 3.4: Demand for digital money......................................................................................43 Figure 3.5: Digital money in the European Union 1996........................................................49 Figure 3.6: Outstanding digital money in the European Union 1996.................................. 57 Figure 5.1: Dynamic behavior for different parameters.........................................................94 Figure 5.2: Increase in money demand.....................................................................................97 Figure 5.3: Stability of equilibrium money balances............................................................. 98 Figure 6.1: The money pyramid...............................................................................................105 Figure 7.1: Private underestimation of risk............................................................................ 149 vi Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Abstract Advances in encryption and computer networks have paved the way for a purely electronic-based currency substitute: digital money. It excels known payment systems in many respects and fits the requirements of Internet commerce uniquely. Many private companies currently try to gain a share of this growing market, allured last not least by the freedom of lacking regulation. The economic behavior of network money differs considerably from paper currency: holdings of digital money will be small. Its introduction comes together with mass speculation and transnationality. Additional systemic risks increase the instability of our payment systems. Technically, also privately supplied money becomes feasible. This amplifies confidence issues and diminishes the central banks’ control over the money supply further. A responsible policy has to recognize the importance of digital money for economic stability and, hence, regulate its issuers. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 In God We Trust - All Others Pay Cash. 1.1 The functions of money What is money? Money is the modem medium of exchange. Media of exchange “can be used in all, or nearly all, transactions and ... extinguish obligations between parties.”1 Because money avoids the inefficiencies of barter and facilitates trade people hold it between exchange transactions. In order to optimally smooth spending over time money might also be held for a short time as a store of value, but other interest paying assets usually dominate money. Furthermore, money can serve as unit of account, the unit in which all prices are expressed, and simplify bookkeeping. If money fails to deliver those services, it loses its attractiveness for the economy. Consumers will then try to substitute it by something else. In high inflation eras, for example, foreign money is often substituted as unit of account because the local money is too unstable to track business activity. Nevertheless, this local money can still serve as medium of exchange. In a credit system the seller needs additional information about the borrower’s credibility to evaluate his ability to repay. In contrast, money already contains this information: It shows that the shopper has the ability to acquire goods and services. Money can be seen as a flexible aggregate record of the past actions, by which money has been earned: It entitles the household to acquire unspecified commodities from 1 [Meyer 1986] p.25 1 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. unspecified sellers at unspecified prices at an unspecified time. Nobody guarantees this future consumption, however, except maybe the issuer or a law. Because money guarantees the ability to repay it makes additional information about the counterpart unnecessary.2 In effect, every market participant has always a positive balance. Therefore, money emulates the immediacy of barter but makes the physical exchange of objects unnecessary.3 On the opposite credit delays the actual value exchange in the future. Whereas a credit payment is temporary and finally settled in something else, money guarantees finality. The linked information costs and the uncertainty about endowments and relative prices affects the credit system. This gives money an added value to possibly dominate credit. 1.2 Money on the balance sheet Money represents liquid value, which is generally accepted in payment for goods or services. On the issuer’s balance sheet outstanding money appears as a liability: If the government or a publicly accepted institution issues money this is usually an interest free loan from the money holder. A bank, however, that creates money by granting credit based on its stock of demand deposits has to pay interest on these deposits. In this sense money can be viewed as a form of credit of small denomination and no fixed maturity. The whole amount of this credit in an economy represents the purchasing power that 2 [Blanchard 1989] p.165 3 [Solomon 1991] p. 188 2 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. finances the purchase of goods.4 Money helps to store, manipulate and account this purchasing power. Lack of it can trigger recessions, but simultaneously credit growth exceeding that of the production in the economy can cause inflation. 1.3 Why has money value? Only in the case of gold and silver coins5 the consumption value of money equals its face value. Today’s “fiat” money6 is not only non-commodity but also unbacked - it lacks any intrinsic value. This money derives its value only from the fact that it can be exchanged, that it makes existing trades easier and that it is scarce. Scarcity keeps its exchange value against goods constant. “If currency is so unlimited in amount as to become practically a free good, people would have so much of it to spend as to bid up all prices, wages, and income sky-high.”7 Only if the issuer limits its supply and its functionality leads to wide acceptance money can have exchange value. We see that people have to believe in lasting scarcity of money to be willing to honor it. The value of money builds entirely on faith, on the faith in that the issuer will keep its supply limited and secure its circulation and acceptance. 4 [Ely 1996] p.l 5 Commodity monies were also shark teeth and stone wheels. Their additional exchange function as monies raised their relative price even further. 6 President Nixon ended the US dollar’s redemption in gold on 8/15/1971. This marked the beginning of the free-floating dollar, but the gold backing for the average person ended de facto already in the 1930s. 7 [Samuelson 1980] p.263 3 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Under competition the value of money and therefore its circulating amount can only vary between narrow boundaries, because otherwise users will discount or even substitute it with other forms of credit. An issuer who cannot convince the public that he will keep his money scarce may want to underline his credibility by the pledge to pay or deliver scare goods and services whenever the money holder wants to redeem his balances. 1.4 The wide variety of money Roughly we can divide money in two categories: inside and outside money. All money that is on net an asset of the private sector is called outside money. The money base, coins, currency and bank reserves belong to this group.8 The government has decreed these tokens to be money, they represent a liability for the government and are therefore net wealth for the private economy. In contrast, inside money is both an asset and a liability of the private sector and therefore no net wealth. This includes all money created by private institutions, especially bank deposits and traveler checks. Deposits subject to checking on demand act like any medium of exchange. Since they posses the same properties they serve as money without any government decree. But they establish at the same time a liability of an issuer in the private sector, usually a bank, and finance its assets. Then, without any restrictions the private sector can create infinite amounts of inside money. Only as long as the 8 [Blanchard 1989] p.193 4 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. convertibility in an easily exchangeable asset like outside money is guaranteed, inside monies will be credibly stable. 1.5 Definitional problems Any commodity can serve as money. Several of them can be circulating in an economy9 at the same time, as long as they are generally accepted for exchange transactions and are storable for at least one period. How much acceptance a commodity must enjoy to be described as money seems to be a matter of degree.1 0 Especially government bonds are highly liquid assets that can be sold for money very easily. A wide definition of money includes these near-money items. 1.6 Private money If a big and publicly recognized institution wishes to do a transaction, it can often execute it without any physical money at all. Only the publicly available information about this institution, that it will finally be able to pay for the purchases, allows it to do business and initiate money-like vaiue transfers. Its reputation has to make up for the lack of reserves; its assets provide the ultimate, but not necessarily liquid backing. A Bill of Exchange works like a money substitute: A trading partner can use this financial instrument to satisfy his creditors, he can easily pass this claim on to other merchants, 9 See [Hayek 1976] p.46 1 0 [Hayek 1976] p.47ff 5 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. without any conventional money flow between him and the purchasing institution ever taking place. Private money can work in a similar way: A publicly recognized private institution, for example a bank, issues private money and lends it to a customer on credit basis. This money does not indicate any real wealth. It merely signals the consumers believe that the bank will keep this money valuable and the issuer’s believe that its agent will be able to pay back his debt. This expectation by the issuer, which is a prerequisite for granting credit and but which does not have to be rational, gives the consumer purchasing power. When and if the settlement of the private money will happen is more a question of monetary policy and system stability than a system immanent necessity. In a perfect economy expensive central bank money would be redundant, because it does not have any functionality on top of the exchange of purchasing power which private money might have. Since the information infrastructure has been growing, more and more private money transactions become imaginable. Private money belongs to the group of inside money. A private banknote can be interpreted as a standardized, abstract form of a credit agreement. Again, its scarcity is crucial for its usefulness as money. Only trust in the private issuer’s actions and his determination to keep the supply of his outstanding credit limited will lead to the acceptance of a private money. The private issuer will have to give a redemption guarantee to signal his credibility, which gives companies with a well-known brand name and a strong balance sheet an advantage. 6 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 Digital money 2.1 A primer on encryption The security of any new electronic value transfer system determines crucially its potential for success. Any implementation of digital money requires not only secure, but also fast and efficient algorithms and data structures for the following operations: a) Verification of coin validity b) Identification of individuals, for example by their signature c) Encryption to protect the value tokens in transit All systems today use cryptography to shield data from interception, duplication and alteration in a public network. I this subchapter I give a short introduction in the underlying cryptographic methods. Let S represent a secret “private key” and P the corresponding publicly revealed “public key”. The message M coded with P can then be denoted as P(M). A wish list for S and P may look like this:1 1 • Each key solves the code that the other key generated: • S(P(M)) = M and P(S(M)) = M • All pairs of S and P are different. • Calculating S from P is very hard and as difficult as decoding P(M) without knowing S. Knowing a person’s P therefore doesn’t help to deduce its S. • Encoding and decoding with known S and P and generating S and P is easy. 1 1 [Schmeck 1996] chapter 3.2.4 7 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. If M is a message that contains digital cash, then the use of P ensures that only the authorized receiver - the person that knows S - is able to read P(M) and therefore only this person can extract any value from P(M). The coding process can also be reversed: If M is a message, for example raw coin data to be signed by the money authority, S(M) is the corresponding certified document. By using P on S(M) everybody can uniquely determine the signer, which is in our case the issuer of the money (= owner of S) and read the content, which is the coin data again. Luckily, keys S and P with the desired properties exist.1 2 The technology underlying digital money is a so-called RSA1 3 cryptosystem, which has become a de facto technical standard.1 4 In contrast to conventional cryptosystems the key doesn’t have to be sent over public channels. As indicated above every agent has two keys. He may publish his public key P in a public database so that everybody can access the information. His private key S, however, is secret and only known to him. One might interpret S as a digital fingerprint. The keys P and S consist of a pair of natural numbers (N,p) respectively (N,s) where N and p are both public and s remains secret. N, p and s are simultaneously chosen by the owner of s or by some trust authority. To do so three big prime numbers s, x, y are randomly generated with s>x, s>y and s<xy=N. p is 1 2 For a more detailed description see [Lynch 1996] p.76f and [Wayner 1996] p25ff. 1 3 RSA stands for its creators Rivest, Shamir and Adleman (1978). 1 4 [Lynch 1996] p.84 - RSA based cryptography is used in the secure sockets layer (SSL) protocol developed by Netscape as well as in the secure electronic transactions (SET) protocol of Visa and MasterCard. See [Loshin 1997] p.77f for details. 8 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. determined by the equation ps mod (x-l)(y-l) = 1 which means that ps has no common factors with (x-l)(y-l). For these numbers the following lemma ensures the desired properties: (mp mod N)s mod N = m^ mod N = m(x 'l)(y‘l)lc+ 1 mod N = = [(m mod N)’(m(x '1 K y ‘l) mod N)k ] mod N = 1 5 = [(m mod N) Tk ] mod N = m mod N = m for all m<N. Depending on the desired security level N consists of at least 200 figures (-512 binaries) whereas s and p contain about 100 figures. Despite these big numbers the necessary modulo calculations are fairly easy thanks to the mathematical properties of our choosing process of p, s and N. The coding process looks quite simple: The software cuts the message M into a series of numbers mi, m2,..., mn smaller than N. Then it calculates q = P(mO = miP mod N. For signing or making coins the only difference is the use of s instead of p. In order to decode the message the computer performs a similar calculation. It splits the incoming data stream again in a series of numbers ci, C 2,..., c„ smaller than N. Then it calculates mi = S(Cj) = C is mod N. To verify the validity of coins we only have to substitute s by the issuer’s public key p. To find out x and y from N and p one would have to factorize N, which is a very time consuming procedure depending strongly on the length o f N. For big N it is actually 1 5 Uses Euler totient function, where m(x '1 X y ‘l) mod xy = 1 holds for all x,y. [Wayner 1996] p.26 9 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. not possible. If the cost for cracking the message exceeds the message’s value enough that any attempt becomes uneconomical, this coding procedure fulfills our requirements. However, technical progress and new algorithms may require an increase in the key length every few years.1 6 Apart from that the discovery of simpler mathematical algorithms that solve the factorization problem could make the whole system obsolete.1 7 If somebody successfully convinces other people to believe that a key he created is the public key of some third party, he can induce other people to send in good believe messages to this third party encoded with this public key. They will never reach their destination, however, because only the person that introduced the faked key can read the message. In order to ensure that no faked public keys are in circulation, a trusted third party publishes them in a key book.1 0 These certificate authorities act like a notary, they check the identity of the signature owner, and are liable for identification errors and therefore reduce the risk of the market participants. In order to avoid conflicts of interest, a government institution would probably be the best solution. 1 6 In 1994 a brute force attack was able to crack a RSA code where N was 129 digits long. See [Loshin 1997] p33. The length of the keys increases steadily: 1024 bits for N are already common and 2048 have also been seen. [Wayner 1996] p.25 1 7 [Wenninger 1996] p. 1 1 1 8 For example Germany’s Signaturgesetz (§4f) and the digital signature law of the state of Utah create such an institution. According to [Wayner 1996] p.81 Apple Computer offers such a system, which requires the user to visit a notary public once. 10 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The described asymmetric encryption operations enable us now to implement the required operations: a) Coin integrity verification: A money authority signs a coin digitally with its private key and merchants and customers can verify the validity o f the coin with the banks public key. b) Identification by digital signatures: Customers sign checks, withdrawals or bank deposits with their private key1 9 and the bank can check the authenticity by using the customers public key. c) Encryption: To send a secure message to an agent, the sender simply encrypts the message with the receiver’s public key. Nobody except for the owner of the private key can decrypt this message. In all, we now have the necessary tools at our disposal to implement a payment system based on digital signals. 2.2 Electronic value Electronic value, sometimes also named “electronic money”, is a generic term for a vast array of new payment mechanisms, which are by no means bound to the Internet. A large portion of the money supply already exists in digital form. Starting with the first electronic funds transfer by Fedwire in the 1960s wire transfers of book entries such as 1 9 Digital signature standard (DSS) 1994 by the National Institute of Standards and Technology (NIST), used i.e. by Verisign Inc. based in Redwood City, California. 11 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. bank account balances now amount to about 90 percent2 0 of the total value of all financial transactions. However, less than every tenth transaction is done electronically, mainly because the big payment networks are only designed for high volume transaction among certified institutions. M l includes most of today’s tracked electronic value but big parts of it are not even measured! Declining costs paired with new technology and an increase in the capacity of computers and networks made the evolution o f a new form of electronic value possible: digital money. This new form of payment tries to resemble conventional money in every way: It still buys physical goods, financial assets, information as well as services. Its spending influences goods’ prices and their volatility. It provides all services conventional money did. But at the same time it is not limited to the conventional idea of money. Issuers can give it additional functionality or pay even interest on it. Digital money has the potential of reshaping our payment systems by establishing an additional open system parallel to the existing bank dominated structures. To clarify the differences and similarities I give a short overview of the newest types of electronic value, many of which are still in the pilot project phase. Roughly we can distinguish between card-based money, which require a physical identification, and network-based money that relies entirely on special software on a computer network. 2 0 [Treasury 19962 ] p.15 2 1 [Crede 1996] p.l 12 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2.3 Card-based digital money Most conventional card-based payment options rely on a credit card: Despite its weaknesses these credit card transactions still remain “the most common payment option”22 in the Internet. They require, however, the transmission of the credit card number to authorize the transaction. This major security gap invites formally to all kind of credit card fraud, as auscultation of the net represents no problem for any Internet freak. Some credit card companies make hacking credit card numbers less attractive by shipping merchandise exclusively to the holder’s home. Others introduced methods to make credit card transactions possible without the need to transmit the credit card number over public channels. Still another approach are clearing systems as the one offered since 1994 by First Virtual Holdings Inc. (FVHI) for information purchases.24 Acting as a message-clearinghouse FVHI can strongly reduce the risk potential and bundle micropayments to one credit card transaction, but all other payment features are the same 2 2 [Madamas 1996] p.78 2 3 Visa and Mastercard International developed SET encryption software to be included in various web browsers. Cybercash also offers secure credit card payment services for the Internet 2 4 This system requires the customer to register in advance at FVHI, which allows him to use a separate personal identification number (PIN) for all purchases, not the actual credit card number. When a transaction is pending, FVHI always contacts the buyer again by email to confirm the purchase. See [SEC 1997] p.5ff and [Loshin 1997] p.l42ff for more details. 13 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. as with a conventional credit card. The bank for international settlements considers this therefore only as a new “access product.”2 5 Digital money solves the deficiencies of the conventional payment systems. Its card-based representative is the smart card, which stores encoded digital cash. When a purchase is made the billing amount is deducted from the card, not from the buyer’s checking account. These reloadable cards separate physically authentication from the communication hardware by entering unchangeable data at the time of production. A smart card contains a microchip, which enables it to store value and to complete transactions on its own in theory without any central database. The usual renunciation of online authorizations comes at the risk of duplication and manipulation of card information. But as most payments will still be done in face to face transactions, this risk may be limited. For payments over a network a personal computer needs a special reading device and software. The card can then be reloaded at this computer or at special automated teller machines.2 6 In addition to its inherent security this card has the potential to not only substitute for the functions of existing prepaid cards2 7 but also to combine financial with non- 2 5 [BIS 1996] p.l 2 6 See i.e. [Wenniger 1995] p.2 2 7 Prepaid cards have been in use in Europe and the US for more than a decade, i.e. universities, telephone cards. Single purpose prepaid cards can be viewed as receipts that allow for later delivery of the good and are therefore not subject to regulation. 14 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. financial applications, for example identification.2 8 Since the microchip on the card can be used for nearly any imaginable task, multifunctionality represents one of the main advantages of the smart card.2 9 In contrast to the conventional closed prepaid card systems, which frequently offer only one possible use, stored value cards may be accepted in many different locations across the country. The most familiar example is the Mondex30 card, which offers no audit trail, but optional PIN protection and up to five currencies on one card. A pocket sized personal card reader makes direct anonymous transfers of card value between individuals possible and displays recent transactions and the value on the card. 2.4 Network-based digital money Network based systems try to emulate their brother payment mechanisms of the real world. Depending on whether their value is account or token based they can be distinguished in check and cash systems. Both rely on sophisticated cryptographic methods and the exclusive use of electronic networks for all parts of the transactions. All electronic value transfers are based on a simple series of strings that has to be authorized by a digital signature. 2 8 Following its main application those cards are also referred to as stored value card or electronic purse. In Europe they are also known as integrated circuit (IC) cards. 291.e. the Visa Cash pilot project at the Olympic Games in Atlanta 1996. According to the Federal Benefits program US Social Security Payments will be paid by a smart card system by 1999. See [Goldfinger 1996] p.7. 15 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Check-type payment systems are currently offered by companies like CyberCash and NetCheque. Both require a bank account, which makes the transactions traceable as the bank collects all the checks again when the receiver tries to cash them in. Check type systems belong to the group o f accounted payment systems. They offer an additional payment medium but do not replace cash, because all transfers can be audited. Since the banking procedure3 1 is the same as with any conventional paper check digital checks are just a new access product, not money.3 2 The representative of digital money among the pure network-based payment systems is digital cash.3 3 Public interest focuses on this last step so far in the development of electronic value transfer systems because it reduces money to a string of digits. Digital cash transfers most of the characteristics of currency in the digital world As with real cash this money requires no intermediary or pre-registration. Nobody except for the issuer in this system has to be certified. Hence, peer-to-peer payments without a third party become possible. The issuer of digital money may be a bank or other trust 3 0 See http://www.rnondex.com. 3 1 The quality sign of a bank ensures the acceptance of this form of payment and increases the security of the system against fraud. Possible fees include charges for the “signature” or “stamp”, with which the bank authorizes that the account owner may start certain transactions - similar to selling a packet of paper checks. Whereas peer- to- peer payments seem to be possible in the long term currently most offers are restricted to transactions between customers and certified merchants. 3 2 There are also hybrid payment systems (i.e. PGX by Systemics Ltd.) that created completely new payment tokens with both cash and credit properties. See [Grigg 1996] p.3. 3 3 Another popular term is “electronic cash”. 16 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. organization - so far it can be anybody - where the customer can obtain it in exchange for conventional money. In contrast to credit cards or electronic checks, digital cash does not necessarily require an online authorization, a link to some third party credit provider or the possession of a bank account. To tackle the danger of double spending each “coin” bears a serial number, which helps the bank in detecting fraud attempts. In all current systems digital money may only be used for one secure transaction. After each transaction the merchant will want to check the validity by contacting the issuer. If the bank does not record to whom which banknotes were issued, it can only confirm that they have not been double spent, but it doesn’t know who used the cash. That way the untraceability feature of real cash is emulated. Narrow-defined digital cash therefore maintains the anonymity of the customer. Digital money targets on payments of rather smaller amounts. Primarily, it replaces banknotes and coins. In the near future the integration of smart card technology and Internet payment systems will further push the popularity of digital cash: Digital money can then be loaded on a smart card and spent independently from the customers’ computer network. 17 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2.5 Definition of digital money As if it wasn’t enough that money has no intrinsic value, digital money also lost its tangibility.34 The proposed digital monies show as main common characteristics that they • represent real purchasing power for which the consumer has paid in advance, • move funds electronically over public networks, • rely on information technology to store and transmit the value tokens, • use modem encryption methods for security. Since the final implementation and success of the digital money systems cannot be predicted yet, all definitions of it have to be very broad to include all the different types. For the scope of this thesis I define digital money as monetary value stored in electronic form on a technical device in the consumer’s possession. It can “be widely used for making payments, without necessarily involving bank accounts in the transaction but acting as a bearer instrument.”35 “This electronic value can be purchased by the consumer and held on the device and is reduced whenever the consumer uses the device to make purchases.”3 6 Digital money includes the flourishing smart card systems as well as the highly publicized but still hardly used digital cash. These new electronic payment mechanisms 3 4 Of course the storing devices like smart cards or hard drives are tangible. 3 5 [EMI 1997] p.30 3 6 [BIS 1996] p. 13 18 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. are standing outside the established monetary system so far and generally require loading value from funds held within the financial system by exchanging deposit or cash for digital value. Digital money is therefore an account-independent claim in digital form on the issuing institution. The key difference between digital and other forms of electronic value is that the former are “supplied with generally accepted purchasing power.”3 7 The generic term electronic money not only consists of digital money but also includes the account-based funds of the existing clearing and payment networks,38 digital checks and other non-cash items. Outstanding digital money balances are still the liability of the issuer, the legal evidence, however, now changes from a piece of paper to a succession of digital signals. The definition of digital money does not require backing funds, but that might be necessary to gain public trust in the value of this form of money Obviously, digital money requires services from financial and non-financial institutions. First, there is the issuer o f money, who guarantees the value and for whom the digital money is a balance sheet liability. There might be only one issuer, probably the government, or several competing private institutions that generate digital money. A money trader may smooth the flow and distribution of digital money by buying and selling value, which shows that the new systems do not necessarily imply mediator-free banking. Network operators and technology companies provide the technical services, for 3 7 [EMI 1994] p.4 19 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. example the cards and software involved. Finally, there has to be a clearing institution for these transactions to settle accounts and to clear issued banknotes. The functions may appear in different combinations, depending on the actual system. Primarily the legal framework will determine the final institutional arrangements. Financial Non-financial I S S U E R Money Trader Clearing House 1 Network operator 1 i ■ C 0 1 N 1 Technology 1 * S 1 Hard-/software i u 1 Provider i M I F , J R Merchant Figure 2.1: Institutional arrangements for digital money 2.6 Example of a digital cash transaction In order to give an imagination of how a digital money transaction works, I briefly describe the approaches of Amsterdam-based DigiCash (Ecash) and the University of Southern California (NetCash) in a simplified form.3 9 A digital coin may 3 8 I.e. Society for Worldwide Interbank Financial Telecommunication (SWIFT). Fedwire, Clearinghouse Interbank Payment System (CHIPS) 3 9 See [Peirce 1995], [Wayner 1996] p.l23ff, 159ff and http://www.digicash.com for details. 20 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. include information like the coin value, a serial number, an expiration date40 and the issuing bank’s name and its server address. By using RSA public key cryptography both systems provide security from message tampering, eavesdropping and theft during transit. The three relevant transaction partners are: • Bank4 1 (B) generates coins, validates existing coins and exchanges “real” cash for Ecash. • Consumer (C) owns account at bank from which he withdraws and deposits Ecash coins. • Merchant (M) accepts coins as payment. Withdrawing digital money requires the following steps: a) C’s computer calculates number and denominations of coins needed. b) C’s computer generates random serial numbers.4 2 In the Ecash system the serial numbers are “blinded” by multiplying the coins with a secret factor.43 4 0 I.e. the “Electronic Monetary System” (Citibank) included an expiration date in the coins. After that day the coins don’t loose their value, but they cannot be used in transactions any more. They must be returned to the bank. 4 1 Note that B doesn’t have to be a financial institution. I use the term ’bank’ just for simplicity. 4 2 A serial number has to be large enough to ensure “uniqueness” (>100 digits). 4 3 The blinding factor k works like this: Instead of sending the plain message m to the bank, the customer sends (mkp) mod N’ where (p’,N’) is the banks public key and (s’,N’) the bank’s private key. The bank signs this: ((mkp) mod N’)q mod N’ = (mq kp q) mod N’ = k(mq mod N’). When the customer gets back the blinded money he simply divides out the binding factor k. 21 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. c) The coins are packed into a message, which C then signs with his private key. Afterwards the signed message is encrypted with B’s public key and sent to B. This ensures that only B can read the message. d) B decodes the message with its private key, checks the signature (with C’s public key) and debits C’s account. Then B authorizes the new coins by signing them with its private key for the desired denomination. Note that in the Netcash system B keeps a record of the issued serial numbers in a database whereas in Ecash B only sees the blinded serial numbers. Since B can’t maintain a database about the real numbers in the Ecash version, it has no possibility to match redeemed coins with the user who got them from B. To protect the coins from theft B encrypts them with C’s public key. e) Finally, the user decodes the message and stores the coins on his local computer until needed. As for Ecash coins he additionally divides out the secret blinding factor. B debits C’s checking account with the appropriate amount in local currency. The Ecash software always keeps enough small denominations to make any payment amount possible. When small denominations get too rare the software asks the user to withdraw additional Ecash to restructure his coin holdings. Note that both trading partners must be willing and able to accept a certain digital cash token, otherwise the transaction cannot succeed. When C hits the webpage of or sends his order per email to a merchant M that accepts digital cash the spending of digital money looks like this: 22 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. a) The software asks C if he really wants to make a payment. Only if he agrees the transaction will happen. b) C’s computer sends the right amount of coins to M, encrypted of course so that only M can read the coins. c) After decoding, M can recognize the coins by using B ’s public key. To verify that they will be honored, he can send the coins immediately to B. He identifies him by signing them with his private key as if endorsing a check. He finally encrypts the coins with B’s public key to secure them in transfer. d) B decodes the message and checks if the coins have been double spent. In the Netcash version B does so by comparing the coins serial number with its database of current coins in circulation. If it appears everything is ok. B deletes this number from the record. In the Ecash version B checks with its database if the coins have ever been spent. If not, the redeemed coin’s serial number is added to the database. If the coins appear to be valid (which means they haven’t been double spent and bear B ’s signature) M is notified about the successful deposit. In the case of double spending the legal owner could prove his ownership by revealing the blinding factor he used. e) M delivers the good. Ecash has therefore the advantage of full anonymity for the buyer but the size o f the database of all spent coins may become unmanageable as the acceptance of Ecash grows. Netcash doesn’t have this scale problem but compromises on anonymity by using 23 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. identified cash.44 Note that keeping a spent list helps to enforce nonrepudiation. M cannot deny that he received the money, because sooner or later he will have to redeem it at a bank, where law enforcement will find the serial number C claims to have spent. Only C can determine by reporting the unblinded serial number if his transactions can be checked. Risks connected to the anonymity of the Internet, which might seduce the buyer not to pay after receiving the good or the seller not to deliver after accepting payment, can for example be solved by introducing a trust agent. By collecting information of both trading partners payment and delivery are locked and synchronized. Including partial identification in each transaction can block potential double spending even more efficiently. The combination of the identification collected in case of double spending would reveal the delinquent’s identity 45 Many other innovations and improvements for digital cash might lie still ahead... 4 4 Anonymity can be achieved by additional exchange possibilities, use of pseudonyms or session specific private/public keys. See for example the PayMe proposal in [Peirce 1995]. 4 5 See [Loshin 1997] p.238f, [Wayner 1996] p.57ff for further improved versions of digital cash regarding off-line spending, fraud-prevention, anonymity, divisibility, etc. 24 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. As the future prospects of digital money are still in doubt, not least due to its economic risks, I want to discuss their various aspects and solutions in this thesis. I will concentrate on token-based digital money systems - smart cards and digital cash - because they are bearer instruments like current currency and contain the most revolutionary concepts. 25 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 3 Necessity of digital money. 3.1 Motivation for digital money 3.1.1 Anonymity Anonymity is the key difference between digital cash and the other electronic forms of payment. Whereas all payment systems besides cash today are accounted and leave the customer helpless to the data collectors, digital cash enables him to do peer-to- peer payments with any merchant and any individual, which increases significantly his ability to do worldwide business from his personal computer. Actually, the consumer does not even need a bank account. This makes a comprehensive tracing of the spending habits as with credit card users, i.e. the amounts, time, locations and type of their shopping, nearly impossible. Some people claim that unaccounted systems are only valuable for those who want to break the law or fear the light of truth. However, many people prefer to keep their business and personal matters private. The acceptance of digital cash compared to other electronic payment systems will show how strong this desire really is. On the opposite side merchants appreciate the ease of collecting market data of all transactions that are not executed anonymously. This makes his micro marketing more effective and more customer-oriented. The transparent customer does not have to become reality, if both accountable and non-accountable digital monies are available and used responsibly. 26 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 3.1.2 Security Credit cards suffer from major security problems, especially when they are used over an open network. In an open system like the Internet the credit card numbers might be tampered in transfer adding to the uncertainty about a merchant’s trustworthiness. Digital money offers a higher degree of built-in security: It ensures message integrity and authentication, since it contains mechanisms to detect any altering or reproduction. Tamperproof smart card technology could reduce the fraud-related expenses of credit card companies by as much as US$ 600 million in the United States and US$ 3 billion worldwide.46 Digital money allows for instant payment when ordering electronic goods and thereby reduces the merchant’s temptation to default on the delivery of the good as known from the credit-card world. Recognizing this problem, credit card companies authorized only selective merchants for card transactions. In contrast, digital money allows also small stores to participate in the market immediately without any previous check. If necessary, digital signatures allow even identifying the trading partner uniquely. The replacement of banknotes and coins by smart cards reduces the incentives for vandalism i.e. on vending machines and - with PIN protection - reduces the consequences of loss or theft. Auditable payment systems also give the law enforcement agencies new tools to track criminal activity, an advantage that is often ignored. 4 6 [DePrince 1997] p.27 27 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. All in all, despite concerns about possible Internet fraud, digital cash provides more security for both customers and merchants than most existing payment instruments. 3.1.3 Lower cost of money The development of digital money makes a new wave of cost saving on transactions possible. Today, handling money costs US$60 billion each year.47 The total payment system expenses for the United States are estimated at between two and three percent of the gross domestic product.48 Since the average cash transaction only amounts to about ten dollars, implied fees make the use of a credit card for them unattractive.49 Table 3.1: Cost per transaction by channel5 0 Channel Cost per transaction in US$ Full service branch 1.07 Telephone 0.54 ATM full service 0.27 PC banking 0.02 Internet WWW 0.01 Electronic payments use existing networks and computers and can therefore be cheaper than any conventional payment form, even credit cards. Credit card fees for both 4 7 [Berentsen 1997] p.8 4 8 [Humphrey 1996] p.935 4 9 Online or small merchants - a group that has a spoiled risk profile fees and are sometimes not even authorized or credit card transactions payment systems. 28 - often have to pay high credit card . They will be the first to adopt new Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. consumers and merchants, especially for payments in foreign currency, are still too high to allow for micropayments.5 1 Only the prospect of nearly zero transaction cost that digital money offers, makes this way of payment for information and other non-priceable goods realistic. In comparison to checks digital money waives the merchant’s risk of a payment being refused due to lack of funds. Also financial institutions profit from the substitution of cash by its digital sister. They now “only lose reserves when the funds embedded in the card are spent and merchants demand payment, whereas, by contrast, when notes and coins are withdrawn by an account holder there is an immediate drain on reserves.”52 3.1.4 Convenience Digital money provides the means to do efficient business on the Internet. But it is by no means limited to the Internet. The interoperability offered by a sophisticated system might allow for downloading it on smart cards or single use cards, i.e. from telephone companies. Its value can be traded and divided almost indefinitely, which ensures its usability for any purpose and any kind of value transfer. It waives the need for small change and guarantee accurate change. Account balances can be updated instantaneously. Digital money has the potential to integrate and simplify payment in 5 0 [Goldfmger 1996] p.2 5 1 See i.e. [Roberds 1997] p.37 5 2 [EMI 1994] p.6 29 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. general. In the global net community there should be monies that are directly accepted worldwide or at least in a large commercial zone to make global shopping easier. High return Low return Figure 3.1: Digital money versus cash Digital money dominates cash with respect to return and convenience but might be dominated for the time being in terms of acceptance and liquidity, because the merchant has to cash in its value. 3.1.5 More efficiency - less friction Digital money increases the transaction speed in two ways: It substitutes for slow cash payments in the real world and makes credit card and similar purchases in the Internet unnecessary. In addition, it makes instantaneous settlement of transactions and a more productive use of funds possible. 30 Bond T-bill Credit card/ Check t Smart card Digital cash I Cash Low acceptance Acceptance, convenience, liquidity Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. But the potential impact on transaction costs goes even further Internet banking requires no banking system with branches and clerks, as we know it today. In addition to the cut in these overhead costs, the low entry barriers of the Internet support an inviting environment for new financial enterprises. Due to the rainbow of technologies involved in Internet banking, the new market participants do not have to be banks. As for today, the most successful actors come from the software industry rather than from the financial sector. Strengthening competition, however, is not linked to national borders. Identical products will have to sell at the same common currency price in different countries. Without any restricting regulation the law of one price dictates banking costs to adopt worldwide. Networks provide total information, all known information can now be reflected in the market. The structure o f the Internet makes even decentralized money exchanges feasible as citizens can easily gain a market overview. If trades between private households become possible the market will be extremely liquid and the money transfer costs - especially across states - will drop dramatically.5 3 When the use of digital money gets more widespread, the cost of exchanging digital monies might become negligible. This will further promote its (global) acceptance and make transnational money flows easier. 5 3 A good example is the success of electronic NASDAQ trading between brokers. 31 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Cost AS(old) less inconvenience, less transaction cost AD(nef) lower cost 11 of money Ti AD(old) v(new) v(old) Volume market expands Figure 3.2: Economic effects of network payments Digital money softens most of the causes o f economic friction like distorting costs, distance, regulations and imperfect information. It makes regional price discrimination more difficult, reduces lags and increases international competition. This is just the next step in a world of decreasing friction. The cost in figure 3.2 include transaction cost, opportunity cost of forgone interest and payment-related risk. If people expect a high risk connected to digital money, they will execute fewer trades. The cost of digital money’s risk is then as distortional as the deficiencies of today’s means of payment. But if the risk of digital money is not substantially higher its cost saving effects shift both the aggregate supply and the aggregate demand curve. On the one hand this increases the propensity of the consumer to buy, on the other hand it expands the supply. Altogether the market grows. 32 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Micropayments allow the pricing of low cost goods, specifically information. When receiving compensation for small parts of information is difficult, it is either under produced or stays unused in a drawer. This slows down economic progress, as much of this information has to be invented or collected several times. Micropayments, therefore, clearly increase total welfare. This digital world meets the efficient market hypothesis more than ever: Current prices reflect all available information so that any future price movements depend on the random and unpredictable arrival of new information. The quick response from consumers can create an unpredictable marketplace, and might cause unwanted side effects such as instability and economic bubbles.54 Altogether, digital money will be a cheaper and more efficient medium of exchange than our current systems. 3.2 Factors determining its success 3.2.1 Government payments The behavior of the government is crucial for the rapid acceptance and the development of standards for the new payment systems. By shifting to electronic 5 4 See also chapters 4.5 and 5.1. 33 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. payments the government can set standards and induce other market participants in the United States to do likewise due to its market power.5 5 3.2.2 Trust Security: Money must be easy to recognize and its validity must be easy to proof. Unambiguously determining the counterparts and issuers is not trivial in a network. A reliable identification system based on trusted third parties and digital signatures can bound the transfer risk. The experience with the transfer of unsecured credit card numbers over the Internet, however, shows that usually not the actual risk but the assumed security determines consumer acceptance. Discussion of the safety o f payment systems or a weak legal situation of customers or merchants will undermine public trust. Consistency and duration are defining properties of money. Whereas an expiration date may be deemed inconvenient, it helps limiting counterfeiting, introducing technological improvements and solving legal uncertainties. The public will also critically watch the behavior in the case o f system crashes and defaults. The conditions of exchange and the timing of the legal transaction must be clear to both parties. Will it be possible to restore the last consistent data distribution? What is the chance o f loosing digital money in case of an interrupted transaction? Privacy: Extensive tracking, marketing and government surveillance has the potential to deter interested consumers. Anonymous systems can diminish the fears of 5 5 The treasury alone directs 800 million US$ annually. [Treasury 19961 ] chapter 2. Furthermore, the debt collection act (1996) shifts all federal government payments to electronic funds transfer by 1999. 34 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. many users to become a customer of glass or their freedom to be restricted by the big brother. Total anonymity requires the possibility to operate without the link to a bank account and to do peer-to-peer payments. The direct conflict with security is obvious. The differing valuation of privacy among agents could create separate markets for digital money. Instant payment: The instantaneous settlement of transactions increases the transaction speed. This reduces settlement risk and cash handling costs by “delivery against payment”. Fast processing is especially crucial for point of sale use, for example at gas stations and kiosks. To build up trust in the new currency also an unproblematic exchange with other currencies is essential. 3.2.3 Convenience Divisibility: Easy exchange of money in sub-units without loss of value is one of the defining characteristics of money. Without it applications like micropayments are impossible. Acceptance: Survey data isolates wide acceptance as the single most important factor for success.5 6 The market participants will only invest in digital cash systems if these base on a common set of standards that ensure compatibility and interoperability. So far few merchants accept digital money, but in order to be useful worldwide acceptance by a wide range of merchants is a prerequisite. A market fragmentation with many incompatible systems destroys most of the convenience digital money has to offer. 5 6 [Panurach 1996] 35 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The household would have to hold more cash than ever - now in digital form — and the foregone interest runs down the attractiveness of the whole system. Furthermore, the length of the innovation cycle influences investors. They will only stop waiting if they don’t expect developments that sweep the existing technology away. Multipurpose and portability: In order to work as a medium of exchange and to imitate money the use of digital money must be independent from any specific physical location or computer network. Storage devices make digital money flexible and allow for the combination with new applications such as identification. Personalized marketing lures merchants by the abundant amounts of survey data, by personalized coupons and automated receipts. Acceptance and multipurpose requires standards. They should not be too tight, however, because otherwise they hinder developments and raise antitrust concerns. User friendliness and reliability: Consumers want convenience. Worldwide web commercial sites are predestined for impulse shopping. Only a fast and convenient way of payment can materialize this excitement for the merchants. The new payment forms make just this possible. Smart cards waive to necessity to carry loose change. The ability to do offline and face-to-face transactions by remote verification without any host connection would give some additional convenience. However, it cannot be repeated often enough that there is a serious security tradeoff by giving up third party authentication. How simple and stable digital money works will strongly determine its popularity. 36 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Fees: Unreasonable high fees are likely to hinder consumers’ and merchants’ acceptance. Nevertheless, according to market surveys consumers should be willing to pay between two and five percent of the transaction value in fees for using the electronic purse.5 7 These fees together with the right to earn interest on the outstanding balances represent the incentive to issue digital money. Since the total amount of circulating digital money will increase with the number of possible applications only increasing competition bounds the issuers’ income potential. 3.3 The demand for digital money 3.3.1 Market environment At the end of 1997 the number of personal computers connected to the Internet totaled 82 million.5 8 Projections for 1999 predict 328 million personal computers, and among them between 120 and 200 million users with direct Internet access.5 9 Currently, the number of Internet users doubles annually and grows at a daily rate of about 0.2 percent60 and could reach one billion worldwide by 2010.6 1 Estimates for sales generated in the World Wide Web range from 2.6 to 13 billion US$ for 1997 62 For the year 2000 5 7 [Wenninger 1995] p.2 5 8 Source: Dataquest press release (August 14, 1997) at http://www.dataquest.com 5 9 Source: [SEC 1997] p.l 6 0 Source: Morgan Stanley survey 6 1 Source: German Department of Economics 6 2 Sources: German Department of Economics, IDC, [Keller 1988] 37 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. analysts see this number between 220 and 600 billion US$.6 3 Two million households in the United States already use electronic financial services with an expected five million by the year 2000.64 As flexible and fast as the Internet, digital money has the potential to evolve as the payment instrument of choice in this expanding marketplace. 3.3.2 Market potential Mainly because of the convenience it offers for small and inexpensive payments cash is still preferred as form of payment. Whereas it is used for more than one in two transactions the total transferred value amounts to only one fifth of all consumer expenditures. In 1994 transactions of less than US$ 10 summed up to more than US$ 200 billion.6 5 Table 3.2: Preferred forms of consumer payment in the USA6 6 Cash 54% Third party credit cards (i.e. Visa) 39% Check 23% Store credit cards 7% Debit cards 1% Other 1% The direct and indirect costs of handling minor cash payments may add up to 15% (for coins up to 40%) of the value of payments6 7 and amount to more than US$ 60 billion 6 3 [Madamas 1996] 6 4 Source: Meridien Research study 1997 6 5 [Congress 1996] 6 6 Source: “Cash-Choice form of Payment" in: Business Week April 8, 1998, p.24 38 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. a year68 in the United States alone. Not surprisingly, the share of electronic payments in the US is growing daily after a period of consumer reluctance. This trend is supported by an increased use of electronic banking, which will be offered by about 300 financial institutions within ten years.6 9 A look at the developments in the United States payment structure reveals some more detail: Table 3.3: Payments in the USA in 19957 0 (Increase 1987 - 1995 in brackets) Value of transactions (trillion US$/year) Number of transactions (billion/year) Electronic transactions '1 544 (+91%) / / 19 (+217%) \ Paper and check transactions \ 73 (+31%) / I 62 (+27%) \ Cash transactions \ 2 (+57%) / / 550 (+97%) ' During only eight years the number of electronic transactions has more than tripled! In the same period the value transferred has only doubled which indicates that the 6 7 Source: Avant Bank http://www.avant.fi/ 6 8 [Panurach 1996] 6 9 Source: German Federal Ministry of Education, Science, Research and Technology 1997 at http://www.iid.de/informationen/iuk/kapitel3b.html 7 0 [Solomon 1991] p.159 and [Solomon 1997] p.39 39 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. average value of an electronic payment is decreasing. At the same time the growth rates of cash transactions are surprisingly high. Reasons for this might be7 1 • Inflation, which causes the need for higher denominations • Additional holdings outside the USA • Growing urbanization (vending machines, etc.) • Expanding underground economy • Population growth • Increased hoarding of currency No matter what reason underlies the increased cash use the prospects for a cash substitute like digital money look bright.7 2 Analysts predict eight billion US$ digital cash circulating in Europe by 2006.7 3 The potential for the United States, if we start with the straightforward assumption that each adult holds $100 in digital cash, is limited to about US$ 20 billion outstanding. That is merely five percent of the total US currency in circulation or about ten to fourteen percent of the currency circulating in the United States.74 At a short-term interest rate of five percent the yearly income for the issuer in this scenario will be five US$ per card. 7 1 See [Moore 1991]. 7 2 See i.e. [Roberts 1997] p.37. 7 3 Source: German Federal Ministry of Education, Science, Research and Technology, 1997. See http://www.iid.de/infonnationen/iulc/kapitel3b.html 7 4 According to the Federal Reserve System, 50%-70% of US coins and currency circulate outside the USA. 40 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. It is at least questionable if this income can cover all marketing and infrastructure expenses plus a profit margin. How people are going to hold their wealth in the future will therefore be decisive for the direction of the development. If only 10% of all households, including many young, high-income households, use Internet banking and these 25 million people hold 10% of their wealth in digital form, this alone would make the exploration of the digital cash market interesting for banks and businesses. “Although their primary purpose it to provide an alternative to notes and coins, electronic purses [and digital money in general] also have the potential to be developed further and to be used for retail payments of greater amounts. In such a case, they would also compete with bank deposits and with instruments that are used to transfer them (such as checks, credit cards and debit cards).”7 5 Indeed, the large holdings of banknotes today are hard to explain. Lacking better explanations, studies conclude that a large amount of currency maybe held as a store of value.7 6 There is no reason why digital money should not profit from the same phenomenon. 3.3.3 A new choice In addition to cash and bank deposits digital money offers the consumer a new choice for his transactions. His choice interacts with and restricts the decisions of the other market participants. The preferences of issuers and sellers may restrict availability, cost or acceptance. 7 5 [EMI 1994] p.2 7 6 [Stuber 1996] p.28 41 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Return d ep osits digital money U2 4 0 - currency ► Acceptance Anonymity Figure 3.3: Concurring advantages The line connecting deposits and currency in figure 3.3 depicts the possible acceptance - return combinations depending on the household’s portfolio choice. The maximum utility it can gain in this situation is Ui. Now we introduce digital money: It dominates cash in return, but it will not be generally accepted for the foreseeable fixture. Nevertheless this new option allows the household to choose any of the portfolios inside the triangle currency - digital money - deposits. The optimizing household will choose a portfolio consisting of only two payment systems — in our example deposits and digital money - and reach a utility of U2>Ui. Although a complete substitution of cash would be feasible, it is probably not optimal in general if we consider not only acceptance but also risk diversification. Even without a legal limit, holding large digital money balances may be unattractive due to security concerns. People can solve this conflict by holding balances for small value 42 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. transactions in untraceable digital money and using only partly anonymous digital money, other electronic payment systems or the good old hard cash for big transactions. Bank Deposits (own accounQ BC, Cash Digital money (third party accounQ Figure 3.4: Demand for digital money Digital money d can substitute large parts of the existing cash C 2 and some interest bearing assets b. Digital cash eases the tradeoff between cash and deposits, a choice between anonymity and security or between convenience and interest income. Especially consumers with cash preference can gain from this new degree of freedom, since they now have a cash-like substitute. 43 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 3.3.4 Precautionary and speculative demand 3.3.4.1 Classical view Following the classical view money demand is proportional to total income Y: M d — = kY . Y can also be seen as the income generated in the Internet. The parameter k depends on the degree of synchronization of receipts and expenditures.7 7 Keynes7 8 distinguishes three motives for holding money: The transactions demand, the precautionary demand and the speculative demand:7 9 3.3.4.2 Speculative demand A speculative demand for digital money appears if it is more stable than other assets. People could flee into the save haven of digital money as a store of value. Another motivation could be risk diversification: They now have the additional choice of distributing the weights in their money portfolio between digital and conventional money. Since neither the one nor the other dominates, the speculative demand for both monies will be greater than zero. 3.3.4.3 Precautionary demand Uncertainty about future spending, interest rates and changes in amount or timing of income cause a precautionary demand. To demonstrate its behavior I use a simple two 7 7 [Meyer 1986] p.460 7 8 [Goppl 1996] p.60 7 9 See also chapter 5.2. 44 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. period model80 with payment at t=l. The probability distribution of the payments ct at time t=l be f(cO. The error we made in estimating the required funds is the random variable e=cp la i,-cr c q u ir e d , where c is the money amount for the transactions. We further define ri as utility loss by illiquidity, for example overdraft interest or waiting cost, and t2 as the opportunity cost of holding money, where ri>r2- With Mp as the precautionary cash holdings for unexpected events the expected utility loss can be written as “ A f p oo E(U(Mp)) = -r{ JCMP + e)f(e)de + rz J{M p + e)f(e)de p If the error e is normally distributed minimization of the utility loss implies that P (e < -M p) = —r -2 — . r{ +r2 The precautionary demand depends therefore crucially on the punishment factors ri and r 2. We get the following results: Table 3.4: The precautionary demand for money P(s<-Mp) Mp* ri=oo or r2 = 0 0 0 0 £ ll H 1 * 0.5 E(e) ri=0 1 0 As long as there are costs connected to occasional illiquidity, the precautionary demand will be positive. Whereas the only cost of illiquidity in a network environment is 8 0 [GoppI 1996] p.73 45 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. the inconvenience of connecting to the bank computer, a lack of sufficient funds on a smart card makes the transaction impossible for the moment. Therefore ri will be higher for smart cards than for network money. 3.3.5 Transaction demand For a formal approach to estimate the transaction demand for digital money I use a simple Baumol-Tobin model.8 1 This model determines the optimal amount of cash- holdings based on a cash-in-advance constraint if we consider only the cost of acquiring the cash and the net of interest lost and utility gained8 2 by holding the cash. The model assumptions are:8 3 • The continuous digital expenditures total Y in the whole observation period. • 2M is the amount retrieved at each bank visit. M are the average cash holdings. • Filling up the smart card or hard disk causes a fixed cost of F. • At an interest rate of r the opportunity cost of holding digital money equals rM/2. Y The total cost of our cash-holding behavior equals C = F h rM . Minimization of C with respect to M implies M = . Optimal holdings of digital money depend therefore on the growth of Internet commerce Y, on the competition 8 1 [Baumol 1952] and [Tobin 1956] 2 M 82 Money offers flexibility and convenience, like in many economic models with money in the utility function. We can also view these services as a kind of interest 8 3 [Goppl 1996] p.60 and [Blanchard 1989] p.l68ff 46 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. among money issuers that influences F and on the interest rate difference between digital money and T-bills r. The decision relevant costs8 4 F for acquiring cash in a network consist o f bank charges and fixed per-transaction costs like exchange fees, telephone costs, time- dependent network usage fees and the opportunity cost of leisure. Grigg8 5 assumes these costs today are presumably close to zero as automation and competition keeps implied cost at a very low level.8 6 Then, as long as holding digital cash has a negative return, the Baumol-Tobin model predicts zero demand. People will acquire exactly the necessary balances prior to each purchase. The optimal behavior changes if digital money is stored on a smart card. As these balances are used for shopping away from a reloading facility the inconvenience and fixed cost of reloading F can be significant. Also in the case of micropayments withdrawing funds continuously may not be optimal, but the optimal cash holdings “will be a trivial amount.”8 7 Providers have therefore an incentive to push the use o f smart cards and to introduce transaction fees in contrast to their current strategy. By holding 8 4 Equipment expenses and monthly fees are sunk costs. Once a bank connection is established also search costs are not decision relevant any more. 8 5 [Grigg 1996] p.7ff 8 6 One of the first market participants, the Mark Twain Bank, charges only exchanges with central bank money, but not transactions that stay inside the Internet world. In the Deutsche Bank Ecash project all transaction are free of charge. 8 7 [Grigg 1996] p.9 47 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. digital cash balances the consumer enjoys having always the exact change, faster transactions and additional features the issuers might offer. When consumers are willing 5 2 5 2 to pay between two and five-percent fee per transaction for this convenience loading fees should be also a feasible approach. In a perfect network banking world money is dominated in rate of return by risk free instruments. The use of small denomination interest-bearing securities as money brings legal difficulties.8 9 Digital money operates essentially in the same way but in protection of a legal gray zone. But holding T-bills for example could be even better; they are too bulky, however, to serve as currency. But if consumers can easily switch between them and digital money without search costs, then theoretically the issuer can induce significantly positive cash holdings only by a positive net of interest lost and utility gained, which means he has to pay interest on money balances or offer additional features. It should be noted, however, that large-denomination banknotes today show an interesting phenomenon: They face a substantial demand although default-free interest bearing assets of similar denomination are available. Probably the anonymity of banknotes and the lack of an easy accessible liquid market for a risk free security give money added value. These properties sustain the demand for digital money and help issuers to skim some of this added value as float.9 0 8 8 [Wenninger 1995] p.2 8 9 See chapter 7.1.2.2. 9 0 See [Grigg 1996] p.10 and [Stuber 1996] p.25 48 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Belgium Denmark Germany Portugal Spain Finland UK Austria EU Figure 3.5: Digital money in the European Union 19969 1 Building on Baumol-Tobin Santomero92 offers a more detailed model for the demand of multiple currencies, but most of his results are ambiguous. In the following I describe the model in a simplified form: A cash-holding strategy has the following profit function: K = rsS + ru M — nu Fu — ncFc — niM Gu — mcGc S and M are average total saving and money holding with their corresponding interest rates r$ and tm. Fm and Fc are costs of exchanging savings into digital money respective currency. Gm and Gc are the transaction cost of the final purchase, n denote the 9 1 Source: [EMI 1996] 9 2 [Santomero 1996] 49 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. number of exchanges, m the number of purchases. If the trips are evenly spaced and the rate of consumption is constant we get especially: M + X u = - ^ ~ and X u = - ^ - 2 nu " 2 mM where Xm are the total purchases performed with digital money and XM denotes average good storage. Following the argument in Santomero9 3 we get after substitution and (rs ru maximization o f 7 c with respect to n M and m M for the optimal nM = J —-— - — — and 2FW r X mM ~ i ~ —~ • Therefore optimal digital money holdings are: V 2G. M M=J±L Fm G A f Kl rs ~ ru V rM J 02 ^ Plugging this result back into the profit function we find — — > 0 , which implies BX M that the profit maxima are either at XM =0 or XM =Y. As long as the combined conditions of digital money (F m, G m, tm) beat the properties currency has to offer (Fc, Go, 0) the household will exclusively use digital money X m=Y .94 9 3 [Santomero 1996] p.948ff 9 4 The interested reader finds the general result in [Santomero 1996] p.952f. 50 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. When digital money dominates, M increases with the interest rate on digital money95 rM and the fee for acquiring digital money FM . To maximize M the fee for doing purchases Gm should be set at a low value, preferably zero. This gives us average digital Y F money holdings of M = ---------- — . With tm=0 this is exactly the result of the Baumol- V 2 rs - r u Tobin model discussed at the beginning of this chapter. 3.4 The supply of digital money 3.4.1 Incentives 3.4.1.1 Avoiding regulation Providers of digital money usually do not evolve from the banking sector. Service providers, payment networks and “brand-name” companies have a comparative advantage to introduce their currency. They can access a wide customer base and influence their payment channels. 3.4.1.2 Float The main motivation for issuers to introduce digital money is float collection.9 6 Issuers can earn interest or investment return from the balances that their customers hold. The outstanding balance provides a form of interest free or low interest debt financing. In addition balances that are not returned because of loss or destruction add to the issuer’s float. The interest income may be low in the beginning if he decides to keep a 100% 9 5 T m will hardly be equal to rs since the issuer has to acquire his funds at a similar rate. 9 6 [Wenninger 1995] p.2 51 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. liquid backing to win confidence and establish his system. With growing Internet commerce the outstanding amount and with it this income is likely to increase. In the long run, however, the float revenues for private companies may decrease again as increasing competition among digital currencies by offering additional services may absorb much of this income. Only a monopoly supplier like the government is able to maximize its revenue by not paying any interest. Without any float income the incentive to issue digital money could diminish. The float is partly paid by the government and the consumers. The latter have to bear it if they substitute digital money for demand deposits. If they substitute it for currency the government has to pay the bill, as it looses seigniorage once it has to redeem national currency in exchange for government bonds. Private money issuers can never q *7 earn seigniorage revenue in the narrow sense, since money balances stay liabilities with unspecified maturity on their balance sheets. Unlike government money they face the legal obligation to redeem them upon demand, they become never owners o f these balances. 3.4.1.3 Marketing effects As we have seen the income from issuing money itself will only give the competitive market return on investments. The main value for the issuer most likely will lie in connected services and public relation effects. First, respected money increases the brand name recognition of the issuing company and works like advertising for the bank’s 9 7 Seigniorage is the difference between the printing cost and the face value of the currency. 52 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. products. Second, u can also be integrated in scrip systems to fragment the market and increase the loyalty of the consumers. Since exchanging digital monies is likely to require the payment of a fee. this could take the grip out of network competition. 3.4.1.4 Financial flexibility Finally, the larger balance sheet gives the issuer additional financial flexibility by gaining an increased borrowing capacity and a strong position in financial markets. If the outstanding balances pay no interest they improve the mix of liabilities. Controlling the unit’s supply by executing the appropriate open market operations he can determine the short-term interest rates for his currency and influence the distribution of his company’s profits and lending revenue over time.98 Furthermore, the short-term-debt property of digital money liabilities proves the commitment of the issuer not to increase the company’s risk. The reduced moral hazard can help the issuer to get better borrowing conditions. Only the long-term stability of the currency measurable in inflation rates, the result of a no Ponzi game condition, limits his freedom. 3.4.2 Costs Issuing digital money creates the following costs: • Setup costs for infrastructure (terminals, servers, cards, etc) • Distribution, collection, network expenses • Marketing 9 8 [Mantonis 1995] p.7 53 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. • Settlement • Asset management 3.4.3 Comparison Simply creating a nice new currency substitute will have no effect if no issuer sees a profit in introducing it. Due to small net cash holdings" the float earned on outstanding balances will be close to zero. That means banks have to rely on fees to earn revenue. However, a fixed charge on balances held does not mirror the true cost structure. People will use the card often and the cost for the providers go out of hand. A charge on the exchange with digital money seems to be workable today1 0 0 but a growing Internet will make frequent exchanges unnecessary. Grigg1 0 1 concludes that per transaction fees will gain dominance. These would, however, destroy the cash character of the new payment systems. 3.5 Current development status 3.5.1 Specific projects Launched in 1993 as a corporate subsidiary of the Bank of Finland, Avant1 0 2 has been developing the anonymous and end-to-end Avant cash-card system. In 1995 its 9 9 See chapter 3.3. 1 0 0 For example the Mark Twain Bank relies on fees from the exchange of conventional currency for digital cash. Long transfer times in exchanges between digital and real money have a similar effect 1 0 1 [Grigg 1996] p. 10 1 0 2 See http://www.avant.fiy 54 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. operations have been sold to Automatia Electronic Purse Ltd., the joint electronic purse company of Finland’s three largest banks. 250.000 electronic purses have already been issued that can be used at about 5000 points of service and 300 companies in Finland. The microchip of the electronic purse rests on the consumer’s bankcard, with a limit on the amount that can be loaded on the card. On certain ATMs he can exchange funds between his bank account and the card for a fee and check recent payment transactions. With a chip card reader for the home personal computer the Avant card enables the consumer to make payments and - starting fall 1998 - also to access public services in some cities over the Internet. Since 1991 the Danmont bank of Denmark has issued more than 375 000 Danmont cards,1 0 3 a prepaid chip card which — like the Avant card - is used for public transportation, parking, public telephones and vending machines. The Mondex card has already been tested in England, Australia and the USA. It is issued by the London-based company Mondex International, which was founded in 1990 and is now owned by MasterCard. In addition Europay and CAFE1 0 4 work on similar electronic purse systems.1 0 5 1 0 3 1996 data, see http://www.pbs.dk/pbs uk/index.html for updates. 1 0 4 Conditional Access For Europe 1 0 5 [EMI 1996] p.684f 55 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. DigiCash is running an Ecash field test with about 25,000 participants and works on pilot projects with various banks.1 0 6 Its currency Ecash has no intrinsic value except for its scarcity and cannot be exchanged in any real currency. Since 1994 one million “cyberbucks” circulate a test market, which can purchase software, documents, services and even “real” goods.1 0 7 The regulators’ approach to digital money so far was limited to 10S several hearings, studies and reform proposals (i.e. reform of regulation E , restriction on issuers). So far less than 100 merchants accept Digicash’s currency. 3.5.2 General trends Most digital monies still are unfortunately in the pilot project phase and the limited market presence destroys the hope for a profitable use at this time. In some countries discussions about the technical implementation delay its success. But nevertheless more and more companies, especially banks, are getting interested in this form of payment as it could revolutionize their core business. 1 0 6 http://www.digicash.nl/. i.e. a pilot project with Deutsche Bank started in summer 1997. 1 0 7 See i.e. [Loshin 1997] p.240. 1 0 8 [Bemkopf 1996] p.4 56 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Belgium Denmark Portugal Spain Finland UK Austria Figure 3.6: Outstanding digital money in the European Union 1996m In Europe besides Denmark and Finland also Austria, Belgium and Portugal took over the lead in digital money. Many other European countries currently mn pilot projects. As seen from figure 3.51 1 0 the outstanding value per card and the average purchase size vary greatly among the countries. The average purchase size among established systems ranges from 98 cents (Finland) to 18.49 dollars (Austria), the outstanding value per card from 90 cents (Austria) to 21.60 dollars (Belgium). For example, the Mondex users load on average 25 pounds on the card and do mostly purchases below 5 pounds. 90 percent of the consumers find the card easy to use - a promising sign.1 1 1 1 0 9 Source: [EMI 1996]. 1 1 0 See chapter 3.3.5. 1 ,1 [Nixon 1996] 57 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. This variety is not surprising. CAFE, DigiCash and others developed a wide range of encryption and features, but the implementation of the final systems is often still undecided. Many offers concentrate on small value transactions, especially on so-called micropayments,1 1 2 and impose a low limit on the maximum value a consumer may hold in digital balances.1 1 3 The lower sums that are at stakes require only limited or no record keeping, which makes a highly decentralized system with full anonymity possible. Digital money value so far has always been denominated in national currency and been issued by a private institution. This doesn’t necessarily have to stay this way, as it is not hard to imagine a wide array of value tokens or central banks starting to issue digital money. 1 1 2 Micropayments are below US$10 and may be even less than one cent to make pricing of “good flows” like CPU time or information goods possible. Its future depends heavily on the acceptance to pay for every single information rather than by one flat rate. It also stands in competition with the established third party payments (advertising). Visa’s Netbill and the CyberCoin service of CyberCash Inc. fall in this category. They differ in the degree of anonymity and security. 1 1 3 The 1997 Ecash pilot project of Deutsche Bank and DigiCash limits balances and transactions to a maximum of DM 400; the Mondex trials in the UK had an upper card limit of GBP 500. 58 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 4 Economic qualities of digital money. 4.1 Security and fairness 4.1.1 Counterfeiting, fraud and theft The Internet is an open network. Since it was created to facilitate the exchange of non-sensitive, in particular non-financial information its openness does not consider the save transfer of value. This opens the door to all kind of security breaks from theft of devices over the creation of fraudulent devices or messages to the alteration of software functions and data, no matter if stored or in transfer. But also disputed charges about delivery and payment can spread if there is not audit trail available. The fraud potential seriously constrains the possibilities of digital money. Since the consumer will not be willing to use any payment system where he has to bear these additional risks, it is likely that system operators and issuers share the responsibility for security breaches. We do not have to visualize the worst case scenario in which the banks private key is disclosed. Already if counterfeiters find a successful way to introduce value in the system for which no payment has been made, for example by duplicating digital cash, the issuing bank can suffer immense losses. In Japan counterfeiting of stored value cards led to losses of about US$ 500 million.1 1 4 1 1 4 [Roberds 1997] p.39 - Certain telephone cards of the German Telekom have also recently been cracked and illegally been reloaded. The loss in this case was at least US$ 35 million. 59 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Especially dangerous in this context are offline transactions as implied by the electronic purse because they make it difficult not only to detect any forgery but also to track the forged cards or tokens. When hiding one’s identity is as easy as on the Internet using an offline payment system results in an environment where fraud can only be detected a long time after it has occurred.1 1 5 If issuers cannot keep counterfeiting at a low level this will destroy the profitability of the whole scheme. Once a security hole has appeared it is unlikely that the issuer can react fast enough to avoid damage. The use of cryptography, a periodic recall and replacement of all coins and a regular update of all security measures seems therefore unavoidable.1 1 6 This should not be surprising as conventional banknotes were also subject to a periodic redesign to keep up with the technological armament of the underworld. These updates will have to be more frequent in today’s network as the speed of innovation has increased. Optimally, the coins and keys should be replaced more often than it takes expectedly to factor N, i.e. to crack the private key of the issuer. The cost of breaking the security walls should exceed the gain by such an attack by several orders. A maximum limit on the amount of storable balances can further limit the potential losses. Even this finds its historic model: The Treasury stopped printing banknotes with denominations above 100 US$1 1 7 last not least for security reasons. Only if the risk can be assessed issuers would find it easy to borrow on their assets and policy could treat forgery as an exception that has not to be taken into 1 1 5 See chapter 7.1.4.3. 1 1 6 [Wenninger 1995] p.3 1 1 7 [Wenninger 1995] p.4 60 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. account in everyday decisions.1 1 8 Whereas forgery attacks can never be ruled out, these security measures require a sophisticated understanding of cryptography and strong computing power on the side o f the criminal to succeed. A more restrictive approach would be to store transaction data, especially for large transactions, on the device to provide an audit trail. These data could be regularly transferred to a central computer, which clears and crosschecks the information. All market participants could use the generated logs to prove their transactions, for example their expenses for tax purposes. Depending on the implementation, however, this could degrade digital money to a mere network-dependent debit system. But we should not be too pessimistic about the security issues: First of all, conventional money could also be duplicated and then double spent. The necessary technology is definitely not high-tech but available to any interested criminal. The average person of course was deterred not only by moral but also the toil of producing a perfect copy. Digital money reduces the hampering to moral scruples and risk of detection; counterfeiting now becomes also suited for white-collars! In contrast to what has been said so far, detecting a crime can become a lot easier since checks with the issuing institution will be more frequent even in unaccounted systems than with conventional currency. Today’s banknotes have to struggle with counterfeiting and everyone who accepts banknotes more or less blindly to deposit them hours later runs into the risk o f accepting faked money. Digital cash can be checked 1 ,8 [Ely 1996] p.2 61 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. immediately for validity even before goods etc. have been exchanged. If the digital tokens are perfect fakes the risk is bome by the issuer. But keeping huge databases and short time lags for counterfeit-detection is costly. Checks are therefore likely to be performed at random. The issuer determines the frequency by weighting the expected losses by counterfeits with the expenses for their detection. This implies that some counterfeits will always circulate, especially among microtransactions, but always connected with a considerable risk of detection. Unfortunately, many countries, among them the United States, have restricted the use of secure cryptography.1 1 9 The motivation is to ensure access to all written communications - an old government privilege, to oversee criminal activity. But the gain in national security is offset by making secure international transactions virtually impossible. Can this government right even be applied to the private keys a bank uses to mint its coins? The approval of the CyberCoin software for worldwide export1 2 0 shows that the US government is flexible if cryptography cannot be used to send secret messages. The safety of stored digital cash and private keys on a computer or smart card is rarely discussed. However, if neglected, this could be the weak point in an otherwise highly protected system. Such essential information must be at least protected by a PIN number to block unauthorized access. Without the user’s knowledge so called “trojan 1 1 9 An individual that uses unbreakable encryption systems is subject to prosecution in the United States. See i.e. [Crede 1996] p.16 f. and Arms Export Control Act (22 U.S.C. 2778) parts 120f. 62 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. horses” and virus programs can intercept the data when it is entered or stored on a local computer. To avoid that the whole transfer protection becomes obsolete digital cash software should rely on strong virus scanning tools.1 2 1 Incomplete transfers may cause economic inconsistencies. In the case of a network crash, the completion of some transactions can be uncertain. Theoretically, both computers could keep backup files of each coin until they have completed the transaction successfully. After a time lag the coins could be removed automatically. This procedure enhances the fault tolerance of the system, but also invites fraud and stands in conflict to the atomicity of money: A transaction should either occur or not. In the case of a transfer interruption here both parties might believe that they own the money token. 4.1.2 Contractual/ legal relationships1 2 2 The missing central control complicates unique identification in the Internet. Today it is easy to introduce new servers or to close them down, the proof of identity requirements are very weak. Domain names or email addresses have therefore no proving power at all. A prerequisite for secure and identified trade in the Internet is the ability of 1 2 0 [Loshin 1997] p. 173 1 2 1 Losses due to careless behavior are usually not covered by the issuer. The Ecash pilot project of Deutsche Bank and DigiCash offers as a more secure storing alternative the possibility to store coins on the banks server and load them online from there upon demand. See §5 of the “Bedingungen fur die Teilnahme von Privatkunden am Pilotvorhaben der Deutschen Bank AG zur Evaluierung des Internet- Zahlungsmittels Ecash”. 1 2 2 See [Artmann 1997] p.l8ff for more details. 63 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. the participants to leave declaratory acts. In written correspondence a signature fulfills this task. If law treats digital signatures the same way they are the most promising analogue for the digital world. A consumer who is negligent with his private key will be as liable for the effects as somebody who is distributing blank signatures. The introduction of digital signatures requires among other related laws the establishment of certification authorities that carefully verify the identity of each signature applicant. Related to the question of declaratory acts is the timing of offer and acceptance. Network commerce blurs the distinction in Anglo-Saxon law between immediate agreements made orally and delayed agreements accomplished by letter and similar devices. 4.1.3 Consumer protection and disclosure To accounted debit systems - systems that transfer value among accounts - the existing laws are applicable.1 2 3 Although only acknowledged companies are allowed to take deposits of conventional money the legal situation for electronic cash is not clear. For the consumer it might be beneficial to restrict issuance to banks as the European Monetary Institute proposed for its members to ensure the customary level of consumer protection. Whereas protection of the consumer for the case of theft and loss of the payment device can be provided by the issuer similar to credit cards,1 2 4 the insolvency risk of the 1 2 3 [Treasury 19961 ] chapter 3 1 2 4 See i.e. [Schreft 1997] p.65f and footnote 7. 64 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. issuer needs a different solution. Electronic money of banks today is only FDIC insured, if the bank stores the value not on the card, but in an account until the payment is made.1 2 5 Money issued by non-banks or even foreign institutions does not even enjoy this deposit protection. If disclose requirements are not strong enough to make all rights and obligations transparent, the difference in treatment between the various systems and also different jurisdictions will lead to confusion and endanger the success of the new technology. For existing electronic debit transactions apply regulation E and the electronic transfer act (EFTA), and for commercial transactions also parts of the uniform commercial code. Regulation E for example regulates issuance, advertising and documentation in electronic funds transfers for all financial institutions. Further, it limits the consumer’s liability for fraudulent use of ATM and debit cards.1 2 7 If downloading value from a checking account by digital money is interpreted as the use of it as an access device this will be protected by regulation E. In shopping transactions, however, digital money serves as cash, therefore suggesting that no protection applies at this time. In a recent revision of regulation E the Federal Reserve exempted stored value cards with a maximum balance of less than US$ 100 and therefore most current applications of digital 1 2 5 [Good 1997] p. 19 1 2 6 See [Treasury 19961 ] appendix 1. 1 2 7 For example, the bank has to proof that its transactions were indeed authorized. Additionally, the consumer is only liable for credit card fraud up to US$ 50. 65 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. money.1 2 8 Without having the right to get a receipt it will be hard for consumers to find out if they have been charged fairly in micro payments and to proof the justification of a claim. In the long term consumer pressure should establish the same level of protection in the network world, as we know it today, but it will take time. All in all this suggests that government should implement statutes that govern the money issuers and transmitters to reduce the risk for the consumer. 4.1.4 Access Providing equal access to the new payment technologies will be a challenge for society. Not only lags the infrastructure in rural areas far behind the economic centers, but also network access is distributed unequally among the population. Countries without sufficient and cheap networks might loose their connection to the beat of the world economy. If big parts of society are excluded form the new payment techniques the idea of a perfect market is vain. The use of cable and satellite data networks can ease the first problem for a transition period and public access maybe granted by community or churches. Nevertheless, households that cannot afford an online connection would give up some control over their money, since they could not use it as flexible as the wired households. Most likely then, they would not accept digital money at all. If, however, the online connection is provided for example by the government, digital money brings an advantage: Since the cost per payment decreases, low value transactions that have been 1 2 8 [Good 1997] p. 17 66 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. uneconomical before may become feasible. This can increase the utility of especially the low-income households. 4.2 (Un-)traceability 4.2.1 Accounted and unaccounted systems The untraceability of transactions in an unaccounted network on the one hand gives ordinary people privacy but on the other hand it makes large scale financial crimes with little detection risk possible. It invites to tax evasion and illegal activities. Digital money systems like stored value cards do not require a name or an address, allow for transfers between individuals and let huge values be spent anonymously in short time. Surveillance of suspects becomes impossible since neither there is any financial document nor does the issuer know his customers. For customers this means that lost cash is irrevocably gone. These cash-like properties of unaccounted systems pose also a risk for the issuer because without records he has no possibility to prove security attacks or to recover and to correct data losses. A possible solution is to require the private agent to keep the records. To ensure that these cannot be tampered the transaction software would automatically generate a representative value (“hash-value”) at a trusted third party.1 2 9 However, if this has to be done for every transaction this is a very expensive if not impossible method. Another approach is to require redemption of the digital money after each transaction as DigiCash does. People can still exchange money among each 1 2 9 [Treasury 19962 ] p.43 67 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. other but then have to bear the counterfeiting risk themselves. That way it loses at least its advantage for the sellers of criminal goods since they can be identified at redemption. Accounted systems keep a complete or partial central record of each transaction. These data are available for audit, which makes tracking of fraudulent or black market activities easier. Being more particular these traces may be even more detailed and offer more control than what could ever have been reached with currency. Accounted systems would also provide an elegant solution for the escheatment laws:1 3 0 Today, balances in inactive bank accounts fall back to the state after same time if no legitimate claimant can be found. Due to the data trail, heirs can be found and dormant funds can be identified more easily. But the other side of the medal is its cost and that many people consider accounted systems as an invasion into their privacy. These systems also lack many of the cash properties, especially anonymity, that a successful substitute ought to have. 4.2.2 Privacy1 3 1 Evil minded persons could use confidential information about somebody else’s financial transactions to destroy his reputation at his creditors or to speculate against his position. Electronic money can give rise to unauthorized access to private information. Data from payment systems on transfer in a network can be intercepted, copied and modified. Whereas the collection of some data, for example spending patterns, can have 1 3 0 Encheatment laws cover the treatment of unredeemed balances. 1 3 1 See i.e. [Artmann 1997] pp,14ff 68 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. positive side effects, collection of information about income or personal attributes definitely constitutes an intrusion in privacy. The opinions about justification of privacy are split: The pro argument in essence points out that every individual has a right to a small space in which it can experiment and learn without having to justify in front of the public’s eye. The contra league emphasizes that strict surveillance helps to make laws and society’s accepted values more realistic.1 3 2 If every little law-disobedience committed with digital money triggers automatic punishment, the distribution of fines is likely to be fairer than now. Either people will start to behave better or they finally abolish the law completely, if people do not support it anyway. Government regulation has to keep the balance between its need for information and the individual’s right to be left alone. For example banks could use accounted systems but promise to deny anyone access to these records. Only with a judge issued warrant the bank would offer access for law enforcement agencies. This extended banking secret could be an area of competition among money issuers and an acceptable compromise between surveillance and privacy: Small offences are shielded by privacy, but if the allegations are severe enough to justify a warrant data is available. Wayner points out: “Most business record in this country are kept secret, but they are still kept. 1 3 2 See [Wayner 1996] p.230 (interview with Stewart Baker). 69 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ... Those who seek to live outside of the country’s institutional memory are usually up to something shady.”1 3 3 4.2.3 Money laundering A money launderer tries to conceal the origin of illegally owned or gained funds (drugs, tax evasion, etc.) to make them look clean. Without the interaction of a system operator funds can be transferred quickly from person to person or to a country with less binding money laundering laws. Government force is difficult to exercise in a network. Whereas big companies are easy to survey, it is the individual that can engage in black market activities giving the authorities no chance to intercept. Digital cash makes further money laundering easier, since it removes the need to transport bulky currency around the globe. The money laundering suppression act of 1994 and the bank secrecy act (BSA) of 1970 regulate commercial money transmitters,1 3 4 but transactions within the private sector remain unchecked. These laws rely on an established financial system, which oversees all data, can filter out suspicious activity and provides this information to the authorities. They do not apply to digital money. But audit trails or limits on the maximum value of digital cash balances could dilute the money launderers’ interest in the new instruments. Furthermore, definitions and reporting requirements for large cash deposits must be updated to include digital money. 1 3 3 [Wayner 1996] p.212 1 3 4 See [Treasury 19961 ] appendix 2. 70 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 4.2.4 Taxation 4.2.4.1 Tax evasion It was easy with credit card transactions: They were fully auditable because there was an independent third party, the credit card company, that kept track of all transactions. If any questions about taxpayers’ expenditures appeared, the IRS simply had to review the records about individuals and amounts involved. As long as electronic money confines to electronic checks nothing will change. With digital money in an unaccounted system, however, the identifiable tax base diminishes as more and more income is earned in networks, where it is irrelevant for tax purposes if the data flows over a public network or an Intranet.1 3 5 Even the possibility to roughly reconstruct the underlying economic activity by assessing the inventory flow disappears if only electronic goods are traded. The Treasury suspects that one main incentive to use unaccounted systems may be indeed tax evasion by underreporting.1 3 6 The sources and character of income, even the definitions of income and money, could become unclear. Wealth can be moved instantly over national borders and flow back as income. The data collection process raises disproportional costs in pure network transactions without any link to the physical and observable economy. The existing tax evasion and administration issues gain even greater explosiveness. 1 3 5 [Treasury 19962 ] p. 16 1 3 6 [Treasury 19962 ] p.38 71 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. In completely anonymous systems citizens will face difficulties to prove their payments without any receipts. However, this can be solved if the taxpayer saves receipts digitally signed by the seller and presents them upon demand. He cannot change them since he does not know the seller’s private key. As solution there is either a redesign of the current tax system or the safeguarding of audit trails, as they are available in the paper world. The Treasury plans to “use the issuers of emoney in this effort”1 3 7 by introducing new audit techniques and reporting requirements, which effectively implies accounted systems with fewer cash properties but forces the evasion potential below today’s levels. 4.2.4.2 Redesigning the current tax system In order for a tax to be neutral, similar income should be treated equal in the network and the real world. A uniform definition of taxable income and uniform and low tax rates is likely to increase the acceptance of the tax. To avoid international double taxation the sales tax should be imposed on the purchase and attributed to the state in which the goods are sold. Without any cooperation among countries monitoring transactions and flows of digital money over the Internet imposes unsolvable problems. Not only micro transactions are hard to trace. The disintermediation realized in the Internet removed all natural “taxing points”, like easy identifiable intermediaries. These financial institutions traditionally cooperated with the tax authorities and obeyed certain reporting 1 3 7 [Treasury 19962 ] p.38 72 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. requirements. All data packages in transfer look the same - no matter if they are letters, money or pictures - and not even the intermediaries are able to control their content. The only possibility is to access one or both of the transaction partners and crosscheck their data for inconsistencies. 4.2A.3 Bit tax A radically different approach would be the introduction of new taxes; the most popular one among them is probably the so-called bit tax. Although the Treasury Department is opposed to any moves in this direction,1 3 8 it would have several advantages. The bit stream can be easily recorded and checked. Network providers are physically linked to a country and therefore easy to identify and access. Finally, such a tax could be socially acceptable, if we agree the heavy information exchange is an indicator for economic strength and that it was the government that made the Internet possible by its start-up investments. However, a bit tax is clearly distortionary, it reduces communication to sub-optimal levels. 4.3 Competition 4.3.1 Low entry barriers Profit margins will be low due to the efficiency of electronic markets. But despite high competition there should still remain enough room to successfully market money and banking products. New market participants enter the market only if the risks of their 1 3 8 [Treasury 19962 ] p.3 73 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. enterprise are calculable, and the rapid market environment of a network can deter many small potential competitors. Non-financial institutions are a much smaller threat to established banks than well-known service providers are, because the formers lack the necessary credibility in the beginning. By bundling digital money offers with additional banking service banks can fragment the market and keep their position in the growing Internet marketplace. 4.3.2 Ease of switching bank and currency As physical locations get irrelevant competition among banks will increase. Induced by competition and low search costs, private money may have additional properties preferred by the household. Governments as monopoly suppliers did not have any incentive to give money more than the necessary design. People might favor money that pays interest among the various providers. Now they can freely choose the best money, which is not only widely accepted but also durable and usable as a store of value. For a money issuer to succeed he has to develop the same attributes we can observe in every competitive environment. Low fees and value added services could build up a loyal customer base. By establishing long-term trust in his currency, and linking this trust to his brand name he might be able to create a profitable niche. If people believe in a currency each note will be longer in circulation and increase the issuer’s interest income. And when other currencies suffer from instability well managed private money units will gain market share. 74 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 4.3.3 Money rating 4.3.3.1 Backing and risk The risk potential of digital money depends crucially on the backing and the ability of the issuer to meet any redemption demand. If short-term default free bonds collateralize completely1 3 9 circulating digital money, then even private money should be essentially risk-free. For non-banks with no access to the discount window a limited liquidity risk remains. More risky and less liquid assets increase the liquidity risk and add a less controllable credit risk. The illiquidity and insolvency risks in a country can be controlled by a central bank that supervises and supports banks in crises. 4.3.3.2 Rating The choice of money is likely to become an investment decision. The question of backing - credit systems versus full backing - will become the basis of competition among issuers. For success issuers will have to actively manage their backing funds. As soon as several monies have gained some importance they will be subject to the rating by information services or rating agencies. Ratings indicate the financial integrity of the issuer and determine the market value of each currency mirrored in a discount off the face value. Comparing the relative rating and pricing of two currencies indicates arbitrage possibilities. Search engines make finding the better money easier. 1 3 9 Japan’s prepaid card law requires 50% of outstanding balances. See [Good 1997] p.22 75 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 4.3.4 Information implications The Internet makes better and more comprehensive comparisons among various offers possible. Sellers have no longer an indelible information advantage. Consumers can get information quickly over networks. Even though today’s free detailed investment information in the Internet might be priced in the future search costs are likely to stay significantly below yesterday’s values. Businesses and households can negotiate deals directly over the Internet. Direct relationships bring the advantage of better conditions for both with them, as there is no intermeditator whose profit expectations have to be satisfied. However, the renunciation on financial intermediators implies that both parties have to bear the full default risk.1 4 0 Finally, the global information network brings the consumer closer to perfect market knowledge. He gets the ability to pick the best offer on the globe, which makes regional price discrimination at the customers’ expense more difficult. Accepting private money requires information. Following the demand the Internet offers more and more monetary and investment information. The conventional time lag caused by financial intermediaries will give place to an immediate response from all market participants. Perfect information might lead to speculation and bubbles. And finally all the offered information may not be viewed1 4 1 because at very low search cost simple time restrictions start to bind. 1 4 0 See existing problems and solutions in forward and futures market 1 4 1 A popular term for this phenomenon is “information overload”. 76 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 4.4 Transnationality 4.4.1 Increase in cross-border payments Increasing Internet commerce will globalize economies further. Since foreign currency can be used for shopping on the Internet the reluctance to hold it will diminish. The evolvement of digital money reduces foreign exchange transaction costs and makes national currencies more convertible. The new telecommunications and computing technologies reduce the impact of national borders and further increase the money flows around the globe in reach of the highest return. As the general public can now easily take part in foreign exchange markets, engage in offshore banking and switch their currency holdings the exchange rates and volumes are likely to be unstable. Cross border payments give raise to an additional risk, the cross-currency settlement risk: During the time lag between the irrevocable outflow of funds in one currency and the receipt of the corresponding foreign currency payment the counterparty can default or the relative values of the currencies might change. Since today the settlement lag can be as much as 14 hours1 4 2 severe losses can pile up. The exposure includes both the principal amount and the cost of replacing the trade. Still, even some states among the European Union have different finality laws and non-overlapping opening times of their clearing systems. The international differences 1 4 2 [Roberds 1993] p.5 77 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. are considerably bigger and another source of systemic risk.1 4 3 However, for the time being the risk added by digital money goes down in front of the US$ 1.2 trillion1 4 4 in net foreign exchange market activity. 4.4.2 Competing national currencies Already today we observe how increasing mobile financial capital in search of global diversification can discipline governments’ monetary and fiscal policies. National currencies and economies are increasingly in competition with one another for financial resources.1 4 5 The information age integrates capital markets further and sets even more mobile capital free. Digital money makes it easy to store different currencies in electronic form. Electronic banking and truly international multi-currency banks diminish the barriers to hold money in foreign countries. More information gives incentives to optimize the money holdings. If one currency shows weakness every ordinary citizen can now shift her savings to a stronger one. In the long run this might drive less stable national currencies out of existence. In the end possibly only the use as legal tender remains; that is to pay for taxes and government fees.1 4 6 1 4 3 The spread of such a shock to other financial institutions in the system is also called the Herstatt risk after a small German bank that went bankrupt in 1974. [EMI 1996] p.675 1 4 4 1995 data. Source: [EMI 1996] p.675 1 4 5 [England 1996] p.2 1 4 6 [Bemkopf 1996] p.6 78 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Worldwide shopping creates a demand for true international currencies. In the long run probably “brand-name” currencies will dominate the network economies. The US dollar or the Euro, for example, have the potential to become a global currency and to substitute local currencies as standard of value. The individual in every country could have a choice among different national money suppliers, and all foreign monies are out of the central bank’s reach let alone control. Policymakers have to consider the reactions of financial markets on their decisions. This competitive game situation poses additional constraints on the policy options, as extravagancies in taxation or inflation result in an immediate flow of capital out of the economy. Compared to the past, government will lose control over their domestic economies. But the competition among national issuers also has a downside: The penalty or utility loss of a government by inflating its currency will no longer be infinite. This causes the same inconsistency problems as among private issuers.1 4 7 But the results are worse: No legal system effectively binds an inflationary government. No lender of last resort will help a central bank once public confidence is lost. The risk of worldwide inflation and depression is therefore not banned, but more likely to grow! If due to regulation only national digital money is competing with each other it will be just a new efficient payment system whose only economic implication is a higher velocity and a changed money multiplier. 147 See 7.3.3.2. 79 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 4.4.3 Supranational legal systems All economic qualities of digital money mentioned above are difficult to manage because of the many different legal, tax and regulatory regimes that are involved. Countries treat the finality of payment as well as the credit and settlement risks differently. What legal system is applicable in cross-border payments will be a primary question in the event of fraud, counterfeiting or default of one participant. If Europe limits digital money issuance to banks and the United States do not, then consequently some US monies should be prohibited in Europe. This makes no sense in the international Internet. Similarly, by low standards in America undermine the high privacy protection in Europe, if European customers shop at a US web site. On the other hand, US law enforcement faces serious problems if it wants to investigate crimes whose roots lead to Europe the same way as it could within the United States. Many compromises will be necessary to build up a consistent hole-free international legal system. Addressing problems before they appear should make a less emotional agreement possible. In any case citizens and foreigners should be treated equally in national laws. An international organization might be necessary to coordinate national laws. If regulations are inconsistent among countries, highly regulated states may observe an exodus of retail banking in laissez-faire countries, because the Internet makes consumer access easy. 80 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 4.4.4 Law enforcement Law enforcement and international taxation are impossible without cooperation, for example in enforcing a sales tax in the Internet. Although the physical location gets completely irrelevant for conducting a business, national laws still end at regional borders. If the legal environment in one country seems inconvenient, it is easy to shift operations to a foreign state, immune from all state jurisdictions. Letterbox companies and trust-accounts containing unreported income can easily be established in a digital money world. It is even possible to operate a computer remotely from abroad without any audit trail, where the national authorities can identify neither the amount nor the citizen to which the foreign income accrues. Information exchange agreements, more detailed than today, are needed to tackle these issues. The progress in fighting money laundering among industrialized nations is reassuring.1 4 8 Tax treaties can reduce the impact of double taxation. That means a country gives up its right to tax the source of income in exchange for having the exclusive right to tax income of its residents. Since physical locations are irrelevant in an international network the traditional source concepts do not work any more. Hence, a simpler approach than concluding tax treaties with every county is giving up source taxation entirely and relying completely on residence based taxation. The residence of individuals can usually 1 4 8 See [Solomon 1997] p.181 81 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. be determined uniquely, but the location of virtual businesses1 4 9 in the legal sense still has to be clarified.1 5 0 4.5 Instability 4.5.1 Speed Rapid moving network money and information require preparedness to act when the first signs of a crisis show up. Contingency plans and institutions with crisis management skills should be established. When all market participants can react instantaneously a daily verification and settlement is not enough because that gives raise to intraday inconsistencies. If the interbank clearing system is not able to follow the speed of market reactions spreading inconsistencies can lead to chaos. Also a time lag in the reaction of the central bank can cause instability.1 5 1 4.5.2 Operational risk Technical risks can never be ruled out completely: The detection of a security flaw in Netscape’s SSL implementation in 1995 shows the versatility of possible mistakes. In this case the cause was as system clock dependent random number generator which made the whole encryption algorithm toothless. This kind of risk can never be 1 4 9 The Treasury for example proposes to view servers that sell information as “data warehouses”. [Treasury 19962 ] p.27 1 5 0 See [Treasury 19962 ] p.22ff for more details. 1 5 1 See chapter 6.3.5. 82 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ignored even after the most careful test.1 5 2 Security holes that allow counterfeiting or fraud pose a risk for system stability.1 5 3 There is no reason to overrely on technology; deliberate use of the new possibilities should always have priority. 4.5.3 Squeezes To avoid the danger of a market squeeze when many users want to redeem their money balances at the same time, the issuer should have the choice among several monies, securities or commodities with the appropriate discounts. Similar to the settlement regulations for Treasury bond futures the statutes should describe a list of accepted ways to redeem the money. 4.5.4 Noise and pre-selection Distributing information and executing trades in a network is much cheaper than over intermediator or conventional direct channels. Waste information can now be distributed cheaply and directly without anybody checking its validity. Information in a big market with many competing currencies will probably be asymmetrically distributed: Big investors get more news — and earlier! Then, regular incompletely informed investors try to make the best out of their situation by following the market leader. Unlike in the past the broad public observes the market and small rumors about a digital money may build up to a big effect. Excessive optimism or pessimism can finally end up in a self- fulfilling prophecy independent of the fundamentals. 1 5 2 [Loshin 1997] p.212f 1 5 3 See i.e. [Solomon 1997] p.224f 83 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Information selection skills become a crucial part in dealing with the flood o f information. Over time information pre-selection companies may evolve. If only a few companies dominate this market, their selection policy will have a strong influence on the market developments. Their advantage by up front information can be limited by regulating their trading activity. 4.5.5 Mass speculation 4.5.5.1 Big market Does massive private speculation really still induce bubbles? If the market participants’ expectations were independent, then an increase in the number of traders would stabilize the markets. The unprecedented number of market participants ensures an arbitrage-free price structure and makes trading large quantities without influencing market prices possible. 4.5.5.2 Bubbles However, the stock exchange crashes in 1987 and 1997 have shown that transactions of the public are dependent and highly speculative. This means that digital money helps the emergence of bubbles. In the past only selected people were directly involved in financial markets. Now lower fees not only make massive participation possible but also reduce the reluctance to trade at even little impulses. This makes speculation even for the common person attractive. This mass speculation will increase cross-border payments and speculation among the various digital monies. 84 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. A huge flow of value out of a certain currency might start a self-fulfilling prophecy.1 5 4 Frankel1 5 5 shows that speculators watch each other more than fundamentals. This behavior can wipe out a private money and start the demise of its issuer. Small issuers face a higher danger of bankruptcy in a speculative attack: Few speculators can bundle enough financial power to return a big part of the outstanding currency. If the issuer cannot keep the price level because he cannot acquire liquid assets as fast as redemption is demanded, the devils circle of more redemption and higher discount has started. Finally, the issuer has to file for bankruptcy. Even for government issued money the cost for society of defending a national currency’s price level might be prohibitive. The result is a huge revaluation. If foreign monies are close substitutes the extent of this revaluation will be similarly big as the effects threatening private monies. Small rumors have bigger effects than in the past. By speculation the small money stock is used much more often in transactions and the money multipliers can be blown up arbitrarily. Every time this money influences market prices, which become more volatile and unstable. 4.5.5.3 Tobin-tax In case of a speculative attack a private digital money issuer still can be bailed out by its central bank until its backing becomes due. Ensuring that this backing is sound contributes to the importance of banking supervision. But because getting international help is usually time-consuming a central bank under attack itself doesn’t have that opportunity. 1 5 4 See models in chapter 5. 85 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. This is exactly why Tobin formulated his idea o f a transaction tax, known as the Tobin tax. One of his main objectives for this tax “is to make exchange rates reflect to a larger degree long-run fundamentals relative to short-range expectations and risks. ... I thought the tax would strengthen the weight of regressive expectations relative to extrapolative expectations. Volatility - in particular detours from fundaments ... - would be diminished.”1 5 6 A Tobin-tax imposes a small tax on every foreign exchange. Due to its low amount it hardly affects long-term investments but makes destabilizing rapid round trips of speculative capital from one currency into the next expensive. The effects of a Tobin- tax differ among the possible implementations of digital money: a) Digital money as a separate currency If digital money is a separate currency, a goods trader will have to do as many foreign exchange transactions as a speculator. As long as the Internet commerce is small compared to the outside economy, an Internet merchant will have to exchange nearly all of his digital money income into “real” money to buy new articles. Furthermore, most implementations require the frequent redemption of the digital cash by the issuer, which is then added to his account denoted in conventional currency. If several established digital monies circulate in the Internet market participants will frequently want to 1 5 5 [Frankel 1996] p.l7f 1 5 6 [Haq 1996] p. xii 86 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. exchange their currency in order to be able to pay at any merchant. In this scenario digital money resembles short-term credit, and will be highly taxed no matter what it is used for. Then, a Tobin-tax artificially increases the usually low demand for digital monies: Just to save the Tobin-tax they would stay in the same currency and try to use the money received for their spending. Renouncing on clearing and settlement they avoid costly exchanges at intermediaries. At most, banks will be able to trade on behalf of the agents, but not on their own account and will construct derivative legal relationships that circumvent the taxing point. Optimally, the agents will form one centralized direct- market to minimize the number of exchanges. In effect, the Tobin-tax involuntarily reduces the central bank’s influence on digital payments, b) Digital money and base money in same currency In this case the redemption of digital money at the issuer does not establish a taxable foreign exchange. As long as economies are not completely interconnected speculators will accrue a higher tax total than investors will. The pros and cons are then essentially the same as discussed in recent publications.1 5 7 Specifically, the increased friction reduces competition among the digital monies. Over short time, issuers have more possibilities in designing their monetary policy, which could introduce a small creeping inflation into the digital money system. Despite these problem areas a Tobin-tax might still be justified in the second scenario: If exchange rates among digital monies show a high variance they introduce 1 5 7 See i.e. [Haq 1996] 87 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. additional unnecessary uncertainty in the economy. But uncertainty leads to underinvestment, as investors value the option of waiting until the uncertainty is resolved.1 5 8 Apart from that additional information costs make the allocation of resources less efficient. The Tobin-tax is justified, if the social cost of these distortions outweigh those induced by the Tobin-tax. The same effects as a Tobin-tax generate required reserves on foreign exchange deposits. Besides the Tobin-tax approach other administrative instruments as in stock market are thinkable, too: circuit breakers, access restrictions and price limits. However, all of them stand in conflict to the decentralized Internet. 4.5.6 Confidence issues Citizens controlled government money indirectly through their vote for certain representatives. This political process not only stabilized the currency policy but also had psychological effect: it strengthened the confidence in conventional cash. Private digital money has to live without this mechanism. It will have to struggle for some time to reach the same trust in the eyes of the public that government monies enjoy because ultimately the citizens stand behind it. Competition and possible bankruptcies will deter risk averse people that do not want to hold a speculative currency. The consumer might rationally consider it too costly to monitor digital money issuers and follow a trigger strategy. This imperfect information blurs the distinction between trustworthy and deceiving banks in peaceful times. Then, a shock such as the 1 5 8 [Haq 1996] p.86 88 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. failure of one issuer for whatever reason can trigger a sudden reevaluation o f digital money’s risk. All digital monies suddenly look riskier. In accordance to the strategy, people dump their digital cash. Doubts about the issuers’ solvency and additional insolvencies lead to a flight into legal tender or other less efficient means of payment. Defaults spread further, until finally even the last person lost its confidence.1 5 9 Historically, often the increased uncertainty induced by the failure of a major financial firm immediately proceeded a financial panic.1 6 0 Such a confidence crisis in the digital world can harm the whole financial system if there is no institution that can smooth the deflection. FDIC insurance and other confidence building measures could lay the necessary foundation to make peoples’ faith in their currency more stable. 4.5.7 Systemic risk Every unsettled transaction contains two kinds of risk: credit risk and liquidity risk. Whereas the order is executed immediately, payment is settled with a time lag, often at the end of the day when the situation of the counter party may have changed: • Credit risk appears when one party accepts orders based on nonexistent funds. If the counterparty never pays the full amount of its obligation the creditor has to bear both the cost of replacing the agreement and the loss of the principal. 1 5 9 See [Eisenbeis 1997]. “° [Eisenbeis 1997] 89 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. • Liquidity risks address the short-term inability of a trading partner to pay. He is unable to raise the funds fast enough to meet the settlement deadlines, but might be able to pay at an unspecified time thereafter. When the risk exposure finally leads to default, then especially if the positions are big a chain reaction of successive defaults can be triggered. This spillover is called systemic risk. Digital monies increase the exposure to systemic risk: Computer programs (“agents”) could continuously optimize the value held in digital currency on hard disk and execute the appropriate foreign exchange or security transactions. Unlike government money, digital money has to cope with the ability of its users to escape from their currency immediately. In the past most people had no feasible possibility to substitute government with a different money even under high inflation. Without any intermediation by banks they can now leave their money of choice immediately at the very first sign of depreciation. Doing so they accelerate the inflation in that currency even further. In absence of reserve requirements the backing of digital money with real currency might be too low to satisfy a run which can eventually lead to the bankruptcy of the issuer.1 6 1 Li a closely linked economy the collapse of one currency can trigger the chained bankruptcies mentioned above and end in a financial crisis. In order to offer favorable conditions, monies might be based on other monies, commodities or even merchandise in storage. However, this strategy not only increases 1 6 1 [Tatsuo 1996] 90 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. the return potential, it also pumps up the interdependence of the payment systems and, hence, the systemic risk due to the many trading partners involved. When new information changes the relative values of the monetary units we observe results similar to the foreign exchange risks of international companies: If a company has income in digital money but expenses in conventional money the devaluation or default of the private money can cause illiquidity.1 6 2 In this environment defaults can spread very fast. As experiences from the free banking period have shown corrective measures by a market manager must respond very quickly to be successful.1 6 3 4.5.8 Real time settlement Current payment and clearing systems frequently built up overdrafts during the day.1 6 4 We have seen that if one part in the system defaults before the effective payment is made at the end of the day this might destroy the whole settlement system. If the clearing system’s reserves are not enough this will trigger an unwinding of payments with impact on many participants. The chance that other participants will get into liquidity problems if they cannot cash in their accounts accruable is considerable. The growth in trade-unrelated financial transactions has increased the settlement risk because there is no good that can secure the trade. Real time settlement among all 1 6 2 See chapter 4.4.1. 1 6 3 [Gilbert 1996] 1 6 4 [Crede 1996] p.5 91 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. market participants will be necessary to avoid a systemic failure of the settlement system. Digital money makes exactly this possible: Its electronic value is a bearer instrument, and this money property ensures payment finality. The implied real time settlement simplifies the web of interlocking loans and liabilities. Even for small firms delivery against payment and a better cash management become possible. Although setting up this system is expensive and no common standards are available1 6 5 so far, the reduced exposure to settlement risk and improved stability of financial markets by shorter, maybe even instant settlement cycles should be worth it. Whereas the risks of trading within one currency will decrease, transactions involving different currencies still remain risky: According to the bank of international settlements “most emoney schemes use existing interbank arrangements”1 6 6 which are sometimes not even accessible at night. These do not seem to be robust enough to handle to dangers connected to settlement time lags in international trading. By the end of 1997, real time gross settlement systems should have been in operation at least in almost all countries of the European Union.1 6 7 1 6 5 [Crede 1996] p.3 1 6 6 [BIS 1996] p.6 1 6 7 [EMI 1997] p. 12 92 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 5 Digital money in macroeconomic models 5.1 The Cagan model 5.1.1 Model setup Digital money is accompanied by a high elasticity e o f money demand with respect to inflation, because it simplifies switching among national or even private digital monies. Not necessarily linked to digital money the speed p at which agents revise their expectations will increase, too. Information is not only conveniently available but also low transactions cost make its instant use feasible. Although Cagan estimated e and p and concluded that the stability condition ep<l is satisfied in most hyperinflations this may not be true any more in a digital world. The basic model based on adaptive expectations can be described by:1 6 8 m = M - = c . e - e * * (1) P V ' dn * — — = p ( K - 7 t* ) (2) at m denotes demand for real money balances, k* the expected and ji the actual rate of inflation; c is a constant. 1 6 8 [Blanchard 1989] p.195 93 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 5.1.2 Inflation If we assume a constant money growth rate a we get from (1): In M - In P = In c — * 1 9 M 13P dn * => cr-n = — — r — = — = ~£p(x ~ x*) M dt P dt dt (3) We can depict this diagrammatically: 169 dn * = 0 dn * = 0 7 C 1 Z Figure 5.1: Dynamic behavior for different parameters Whether constant money growth a causes a hyperinflation in this model depends on the product ep. If ep>l then an external shock can cause a hyperinflation in this currency. If inflation is higher than expected people increase their inflation expectations too fast, hence reduce their money demand. Thereby the price level rises and inflation reaches even higher levels. 169 [Blanchard 1989] p. 197 94 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Unfortunately, both the speed of this process as well as the demand elasticity for digital money are higher than the current values so that hyperinflations in these currencies are more probable than in the national currencies we observe today. 5.1.3 Seigniorage The higher elasticity reduces the seigniorage S obtainable: „ dMldt dM/dt M S = ------— = --------------= om (4) P M P In steady state 7t*=a, therefore we get with (1) S = o c-e~ea . Maximization of S with respect to c implies a revenue-maximizing money growth rate o f < 7 = — .1 7 0 £ However, digital money demand has a higher elasticity 8dm with respect to inflation than conventional money Sc- People switch faster out of a specific digital money 1 c 1 c and the issuer can therefore obtain less seigniorge SD M = ---------< Sc = -------than today’s ^ d m e £ c e government. If better alternatives are available 8dm should be close to infinity in perfect competition, making seigniorage collection impossible. 5.2 A Sidrauski model 5.2.1 Model setup Digital monies offer added features and more convenience. The marginal utility of having more money should therefore be higher for digital money at every money-holding 95 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. level m. The Sidrauski model puts money in the utility function and assumes perfect foresight (jt=7t*). The household preferences can be described by: oo maxU = J u(ct,mt)e~* dt subject to its budget constraint c, + ~j~ = y ~ Ktmt c>0 denotes consumption, m>0 real money balances and y the income in every period, each of them per capita. The first order conditions of the Hamiltonian H = are \ U C (ct,mt) = // Un(ct,mt) = i i t7C'+fite + iil ]hn{ite~amt = 0 (5) 1 7 0 [Blanchard 1989] p.198 96 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. If we assume separability of U Uc(y,mt) is constant and in equilibrium1 7 1 /r = 0. U„ (m,) Hence — — — = k +Q . U c ( y ) 5.2.2 Money demand By choosing units so that Uc(y)=l we can write U +6 . We can express the money demand therefore by mt =U^{7U + d) (6) m Figure 5.2: Increase in money demand Since digital money increases the marginal utility Um of money in general for all m the demand for money grows, too. 1 7 1 [Blanchard 1989] p.243 97 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 5.2.3 Stability We can define the rate of growth of real money balances as the difference between the rate of growth o f nominal money and inflation dm (6 ) =>-----= (< T — 7t)m =(<T + 6)m — Um (m)m dt m (7) If the issuer does not something weird we can assume that a is a constant. The diagram shows two paths: dm /dt A Path A Path B Figure 5.3: Stability of equilibrium money balances Hyperdeflation represented by path A can be ruled out since it would violate the transversality condition (5). Hyperinflation in path B, however, can only be ruled out if 1 7 2 See equation (3). 98 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. lim mC/_(m) > 0: Money must be so important, that the marginal utility of money m-»0 increases faster than the real money stock goes to zero. If there are alternatives like foreign digital monies or private issuances, they can easily be substituted. A renunciation of money A combined with a shift to another equally good money does not change the total utility at all. U ^ a for a specific money A is small as it is bounded by the utility other second-best monies can offer. The steady state M then may be not only unstable but also not unique. The possibility of — — » 0 makes self driving inflations in specific competing monies more probable than in the past. 5.2.4 Credibility The convergence to zero real balances can be blocked if the issuer or any kind of institution guarantees a low fixed redemption in any situation. Then, if inflation was gaining speed and the price level was growing, the value of money would find its lower bound at that guaranteed price. Since people get this value for sure, they have no reason the follow further speculation. Essentially a feedback rule has the same effect: The issuer reduces a when the price level gets too high, for example < J — Um (m) —— m => = (Um (m) — — + 6)m — U m (m)m — dm — a. dt m 1 7 3 [Blanchard 1989] p.241 99 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Both rules ensure by limmUm(m) > 0 that zero real balances cannot be a feasible m->0 solution. To avoid speculators testing the strength of the promise both commitments have to be credible. “Feasibility of the governments [or any issuer’s] policy requires that the government have access to sufficient reserves o f capital to purchase the entire money stock M at the support price. ... Otherwise, speculative inflationary paths involving a sudden run on the government’s reserves cannot be excluded.”1 7 4 However, “even the possibility of intervention suffices to rule out speculative paths.”1 7 5 For private issuers this will be difficult to accomplish on their own. Institutional arrangements are necessary: Government laws or an obligatory insurance of circulating balances are commitment technologies that can build up credibility. 5.3 Other models The cash-in-advance constraint is still valid in a digital money environment as can be seen from the technical implementation:1 7 6 Money must be used in most transactions. The demand for money per period equals the planned and actual purchases per period. Problematic, however, is the assumption that the velocity of money equals one period. Especially network money can be acquired instantaneously so the period between two trading intervals can be arbitrarily small. A pure cash-in-advance model therefore does not reflect the reality. 1 7 4 [Obstfeld 1983] p.684 1 7 5 [Obstfeld 1983] p.685 1 7 6 See chapter 2. 100 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Models that simulate credibility offer more insight. Government money issuers have to plan with an infinite time horizon. If government loses its reputation it will suffer forever, since it can change its identity only with difficulty.1 7 7 In contrary, private issuers with limited liability have a termination option. Later, investors can refound a new company with an untouched reputation. If there exists a non-private alternative, the private money system as a whole might suffer, but the individual issuer will not account fully for this cost. Whereas competition may not prevent default of private issuers, it makes inflation unattractive, since this will cause a flight in the competitors’ 178 currencies. I discuss these models in the chapter 7 under the corresponding institutional arrangements. 1 7 7 See chapter 7.3.3.2. 1 7 8 See chapter 7.1.4.3. 101 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 6 Impact on central bank policy 6.1 Money supply 6.1.1 The monetary aggregates 6.1.1.1 Current measures Following the many different forms of money that circulate in our economy, measuring its aggregate amount is not an easy task. The Federal Reserve System defines four different point-in-time categories1 7 9 to measure the money stock at the end of each day, dependent on their relative liquidity. M l1 8 0 tries to measure money used as medium of exchange. It contains currency, travelers’ checks, demand and other checkable deposits. Generally, currency amounts to 1 7 9 Composition of the money stock measures and debt as defined in the Federal Reserve Bulletin, March 1998, Board of governors of the Federal Reserve System, Washington, DC: 1 8 0 Ml consists of (1) Currency outside the U.S. Treasury, Federal Reserve Banks, and the vaults of depository institutions; (2) Travelers’ checks of non-bank issuers; (3) Demand deposits at all commercial banks other than those owed to depository institutions, the U.S. government, and foreign banks and official institutions, less cash items in the process of collection and Federal Reserve float; and (4) Other checkable deposits (OCDs), consisting of negotiable order of withdrawal (NOW) and automatic transfer service (ATS) accounts at depository institutions, credit union share draft accounts and demand deposits at thrift institutions. 102 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. only a small part of M l, especially in counties with heavy check use like the United States.1 8 1 M21 8 2 tracks the use of money as temporary store of purchasing power. It includes therefore not only currency and checkable deposits but also all highly liquid assets that are held to bridge the time gap between cash in- and outflow. These assets can be quickly exchanged into Ml money without any loss in value. It consists of M l plus • saving, • small-denomination time deposits, and • balances in retail money market mutual funds. 1 8 1 See i.e. [Champbell 1988] p.27 1 8 2 M2 consists of Ml plus (1) Saving (including money market deposit accounts MMDAs), (2) Small- denomination time deposits (time deposits-including retail RPs-in amounts of less than $100,000), and (3) Balances in retail money market mutual funds (money funds with minimum initial investment of less than $50000). Excludes individual retirement account (IRAs) and Keogh balances at depository institutions and money market funds. 103 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. M 31 8 3 includes less liquid assets, which are usually not exchangeable into cash immediately. However, they are risk-free with only a short remaining time to maturity. M3 expands M2 by • large-denomination time deposits, • balances in institutional money funds, • RP liabilities, and • Eurodollars held by U.S. residents at foreign branches of U.S. banks worldwide. L is a more comprehensive measure. It tracks the use of the dollar as a unit of account by inclusion of all items that are directly based on the dollar standard. L includes in addition to M3 the non-bank public holdings of U.S. savings bonds, short-term Treasury securities, commercial paper and bankers acceptances and the net of money market mutual fund holdings of these assets. Separately measured, the debt aggregate1 8 4 is the outstanding credit market debt of the domestic non-financial sectors. Among other loans it includes mortgages, tax- exempt and corporate bonds, consumer credit, bank loans and commercial paper. 1 8 3 M3 consists of M2 plus (1) large-denomination time deposits (in amounts of $100,000 or more), (2) balances in institutional money funds (money funds with minimum initial investment of $50000 or more), (3) RP liabilities (overnight and term) issued by all depository institutions, and (4) Eurodollars (overnight and term) held by U.S. residents at foreign branches of U.S. banks worldwide and at all banking offices in the United Kingdom and Canada. Excludes amounts held by depository institutions, the U.S. government, money funds, and foreign banks and official institutions. 104 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Reserves, vault cash and currency outside of banks represent the monetary base B, with which all other M’s are connected by the multiplier m. 6.1.1.2 New definitions The digital money is further away from the money stock than anything before: Digital money and other payment innovations without link to central bank system Electronic money with linktajj central bank system ^ Conventional “M” mohe? 1 Private money Central money J Figure 6.1: The money pyramid 135 1 8 4 The debt aggregate is the outstanding credit market debt of the domestic non-financial sectors - the federal sector (U.S. government, not including government-sponsored enterprises or federally related mortgage pools) and the nonfederal sectors (state and local governments, households and nonprofit organizations, non-financial corporate and non-farm non-corporate businesses, and farms). Nonfederal debt consists of mortgages, tax-exempt and corporate bonds, consumer credit, bank loans, commercial paper, and other loans. 1 8 5 [Solomon 1991] p. 17 105 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Electronic money has no time limit. Only at the time of final settlement it merges back into the conventional money stock M. Because it is more or less regularly settled in deposits or cash a weakened influence of the central bank’s policies and an indirect backing by the conventional reserves still applies. Since the data is published irregularly, monetary policy faces its first data collection problem here. Digital money causes much bigger problems. Although usually finally settled in real money the impact of monetary policy is weakened by the layers in between. People can trade infinite amounts o f digital money among each other and issuers can even settle them with different assets. As other financial innovations outside the M’s evolve the old M aggregates lose successively their explanatory power. In order to maintain their information content they should be expanded to electronic and digital money products. The switch of the Federal Reserve Bank from M l to broader measures like M2 and M3 in 19821 8 6 as well as the successively broader definitions of M2 in the past reflects the growing variety of monetary products. Further updates in all monetary aggregates will be necessary. This is especially true for the currency aggregate Ml: Digital money is a perfect substitute for notes and coins and to some degree for sight deposits. Theoretically, people could just substitute current money by its digital sister. Therefore M l should include not only physical cash and deposits but also information-based money flows like digital money. 1 8 6 [Solomon 1997] p. 102 106 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The measurement of the M aggregates only looks at their end-of-day values, not at their dynamics during the day. Solomon1 8 7 proposes new continuos measures for electronic based money, its flow and velocity. It is the gross payments flows that determine the amount of money with demand effect. Since electronic trading has reached unseen volumes a change to flow measurement techniques will have a big impact on measured values.1 8 8 6.1.2 The behavior of the money aggregates 6.1.2.1 Mechanisms a) No binding reserve requirements Banks often hold reserves in excess of what is required by law for settlement purposes. Insufficient reserves could imply costly overdrafts or even illiquidity. We see there are many reasons for non-binding reserve requirements. This is exactly the situation we observe currently in the United States. The Federal Reserve Bank supplies passively whatever quantity of currency the economy wants to hold. Also, by following the business cycle induced demand for money it always supplies sufficient reserves that the banking system can meet its requirements. It even pays interest on the reserves that can only be used to pay for central bank services. “The reservable deposit portion of the money supply is entirely demand driven.”1 8 9 1 8 7 [Solomon 1991] p.32f 1 8 8 [Solomon 1991] p.37 1 8 9 [Ely 1996] p.3 107 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Required reserves in such a case do not limit the quantity of reservable deposits in the economy, but the price of them due to the foregone interest. The Federal Open Market Committee (FOMC) relies solely on interest rate signaling to influence the overnight lending rate on Federal Funds and simultaneously the amount of excess reserves. The Federal Reserve can only move in a very limited policy window because otherwise banks might dissolve their excess reserves and borrow somewhere else. If banks use the Federal Funds market only sparingly the central bank would have to go back to binding reserve requirements to regain its influence. When there are excess reserves ER a change in them by a cash withdrawal has no implications for the bank. The marginal cost of deposits is equal to the marginal return on loans and digital cash. The market for deposits and loans is in equilibrium and the bank could therefore be indifferent between conventional currency and digital cash transactions. However, the exchange of national currency in digital cash reduces the currency in circulation. After having deposited a possible reserve requirement on digital money balances tdm, the bank can use (l-rpM) of it either to issue credit or to increase it excess reserve position ER,1 9 0 which would lower the bank’s risk of having to borrow short term for settlement purposes. 1 9 0 The banks can earn interest this way. They could also hold the deposited currency as vault cash. I ignore this possibility here since the banks would lose the interest that way. 108 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. This oversupply of reserves and other interbank credit lowers their price, the Federal Funds rate. Therefore more loans become profitable, the aggregate outstanding debt in the economy D rises, only bounded by the amount at which the reserve requirement rc would be binding again: D < RR/rc. How strongly D will rise depends on the credit demand function of course. It determines the effect on M l. If digital money is included in M l, we observe a clear expansion in M1=C+D+DM, because the change in C completely offsets the change in DM, but D is growing. If we do not include digital money1 9 1 in M1=C+D, the result is ambiguous depending on the dynamic of D. In general we can expect a multiple effect on M l like in an expansionary open market operation: The additional liquidity is provided by the digital money issuers. b) Binding reserve requirements A bank without excess reserves has to acquire new reserves for every new loan. Binding reserve requirements in the whole banking system mean that it is impossible to get unused additional funds from somewhere else. The bank can therefore only grant a loan if its return exceeds the cost of the corresponding deposits plus the cost of retiring money in reserves. This implies that in equilibrium the return on loans exceeds the cost of the deposits. The spread is unrealized profit due to the reserve regulations. If reserve requirements on national currency are binding and reserve requirements on digital 1 9 1 A reason for not including digital money in the monetary aggregates like M l might be that it has a different denomination from the national currency. 109 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. currency are not the bank has already an incentive to switch to the latter just to escape this reserve tax.1 9 2 A similar incentive exists today for private parties to borrow and lend directly without an intermediary. If the reserve requirements are binding, a currency withdrawal reduces the bank’s reserves by the withdrawn amount, the required reserves, however, decrease only by the reserve requirement quota. The bank will therefore have to obtain additional reserves in a conventional money transaction. When the currency is deposited again, excess reserves give the bank new lending scope. In contrast, handing out or cashing in digital cash only changes the bank’s demand deposit liabilities and if the reserve requirements on demand deposits and digital cash are equal, the reserve position does not change. We can apply the multiplier model: The effective amount of M l is equal to the multiplier m times the monetary base B. The required reserves are RR = rcD+roviDM, where D are demand deposits and DM are digital money balances. The monetary base can be expressed by B = C + RR + ER, where ER are excess reserves. Assume that changes in currency and digital money are related by dC=k-dDM. If people simply substitute digital money for currency (dC=-dDM) k is equal to -1. If people add digital money to their current currency holdings although the central bank did not increase the monetary base k will be 0. In the case of simple substitution an exchange of national currency C in digital money DM increases the bank’s digital money liabilities and holdings of central bank 1 9 2 See [Wenninger 1995] p.3 110 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. money. The banks reserve position1 9 3 RR+ER increases by C but it only needs td m C to satisfy the reserve requirement on digital cash. It can therefore use (l-rDM)C to build up excess reserves ER and directly expand lending: -dC — rcdD + rovidDM + dER (*) This is lucrative because binding reserve requirements imply that the return on loans exceeds the cost of deposits. As banks increase lending this return on loans is likely to decrease. If we assume that this return nevertheless always stays above the cost of deposits, the banks will always want to lend more because the return is higher than on excess reserves. If we include as suggested digital money we can define M l asMl = C + D + DM. Total differentiation implies dMl = dC+dD+dDM. Assuming that the excess reserves do not change (dER = 0) we get with dC = k-dDM from (*) dD = — ^+ rD M dDM and from r c for ~ ~ m Jc “ v / * that dMl = — ---------— -dDM . A rule to induce neutrality of digital money on M l rc is to ensure td m = (l+k)rc-k. For k=-l this is essentially a 100-percent reserve requirement. Otherwise digital money causes either an expansion or contraction of M l. The following table sums up the effect of digital money on Ml: 1 9 3 The banks could also hold the deposited currency as vault cash. I ignore this possibility here since the bank earns no interest that way. I l l Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 6.1: Impact of digital money on an expanded M l1 9 4 d M l d D M Scenario efficient dC = -2DM substitute dC = -dDM additive dC = 0 complement dC = dDM general dC = k-dM k = -2 k = -l ** I I o k = 1 tdm- 0 2 - r c — >0 r c 1 2rc —1 k r c — k + rc ■ - - ^ \J rc rc rc tdm = rc 2 - 2 rc --------—>0 r c V o 0 0 V t # 1 k l ~ r' rc tdm = 1 1 — ^ n i V o 0 A © 2rc —2 „ —£-----<0 rc 1 + 9 I general 2 r° r°- M - >0 _ 1 - V > o rC rD M 2rc ~ rD M ~ ^ krc k + r c rDU rc rc r c r c r c As expected M l increases most likely if there are fewer restrictions on digital money than on conventional money td m ^ c- A less regulated digital money exchange works like a relaxation of reserve requirements, effectively reducing the cost of deposits. Note that even if tdm=0 the increase is only in the additive case unbounded: From k=0 follows dC=0 and rcdD+roMdDM=0. Therefore dDM = —— dD and r D M t — / * p r d M l=-££— — dD = —— — dDM . As long as tdm>0 an increase in DM requires a r D M r c reduction in demand deposits D, the total increase in M l is therefore bounded. In the case of roM=0, however, M l can increase without limit. 1 9 4 Note for comparison that the deposit of dC into a checking account dD gives ^ _ _L _ _ dD rc 112 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Even if the reserve requirements were binding in some economies at the moment the raise of digital money could change this soon. As people need less central bank currency and banks find ways to settle their overdrafts in the capital market, the demand for currency and clearing balances drops and relaxes any existing binding reserve requirement. Binding reserve requirements have an additional implication: Growing economic activity in the networks increases the demand for digital money. And the internet-related services’ share of in the gross national product is indeed growing. If the regulatory framework causes a shortage of digital money the interest rate on this kind of balances will rise, inducing a capital inflow and a reallocation of resources between the digital and real world. Independent of its attitude towards digital money the central bank cannot ignore the effects of increasing electronic commerce. 6.1.2.2 Disturbances a) Digital money as reserve The distortion potential of digital money looked limited in table 6.1, its behavior resembles that of common demand deposits. The calculation above assumed, that (other) digital monies could not be used as reserves for issuing new money. But if a digital money A can be used as reserve for another digital money B M l can increase infinitely without any link to central bank policy. The created digital money can be reused to issue even more of it. Assume now that digital money can serve as a reserve for deposits only and that deposits as today cannot be used as reserve for any kind of money. Then, one dollar of 113 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 base money can create digital money balances. They can be deposited again: One digital dollar again can at the maximum build up — dollars of deposits. Therefore M l The result reflects the known effect of fractional banking increased by the new possibilities digital money offers. If there are no new possibilities, which means a 100% reserve requirement on digital money roM=l. the introduction of digital money does not alter the M l multiplier. In all other cases digital money as reserve undermines rc. If one of the reserve requirements is set to zero the money multiplier and Ml have no upper bound. If as in some countries there is no reserve requirement on demand deposits rc=0, then using digital money as reserve cannot worsen anything. If deposits and digital cash are subject to the same reserve requirement r, the amount of M l that can be created on top of the monetary base B can be expressed by — B . As long as the requirement r is positive infinite money creation can be ruled out. If the central bank wants to keep the money multiplier unchanged it could choose a new common required reserve ratio b) Expansion of Ml Depending on their implementation digital money products can change the demand for all parts of central bank currency. Substitution by digital cash reduces not only the stock of central bank currency; it also changes reserve positions and the size of q » n can be expressed by M 1 = B • ^ ^ (1 - rD H )k (1 — r c )"“* n B n=0k=0 114 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. deposits. Whereas all monetary aggregates will be affected, the strongest impact will be on M l since the comparative advantages of digital cash are biggest in this sector. At first M2 and M3 are affected only indirectly by M l, but the evolvement of products that use digital money as underlying asset could boost the change in M2 and M3 to a level similar to M l. The impact of digital money on the measured monetary aggregates, of course, depends strongly on their definitions: If M l includes digital cash, the public substitution among monies happens completely inside M l. If influences the value o f M l no more than bank deposits. Berentsen1 9 5 gives an upper limit of the possible expansion of M l: For binding , i . . . . 3Afl C dM 1 c reserve requirements the elasticity of M l can be written as e = —--------- = -------------- , dC M l dC 1 + c where c denotes the currency to deposit ratio C/D. The elasticity gives the effect of a one- percent change in C on M l. For roM =0» substitution (dC=-dDM) we get from table 1 C 1 c 6. 1.£ = = . rc Ml rc \ + c Results for various countries can be obtained from the following table: 1 9 5 [Berentsen 1997] p.7 115 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 6.2: M l elasticity for various countries1 9 6 c rc e Austria 0.60 0.050 8 France 0.18 0.010 15 Germany 0.42 0.020 15 Italy 0.19 0.150 1 Japan 0.37 0.013 21 Switzerland 0.44 0.035 12 United States 0.45 0.100 3 c) Foreign denominations The question of whether digital money is a new currency or still part of the old national one might be superfluous. If electronic money can be spent in several currencies instantly without any formal exchange and is issued by domestic and foreign sources measurement problems appear in any case. It is unclear what part of these international balances should be included accurately in the measurement of the digital money in home country currency. Even worse, digital money - for example denominated in US$ - could be issued abroad with a redemption guarantee in paper US$ but no reserves or US jurisdiction. If this kind of money became successful it would establish a money completely out of the 1 9 6 Sources: [Berentsen 1997] p.7 and [BIS 1996] p~5, IMF and national central banks. Canada and the United Kingdom have no reserve requirement. Canada requires non-negative average settlement balances before overdrafts; the UK uses a cash-deposit ratio. 1 9 7 Reserve requirements on transaction and sight deposits. Since the reserve requirements are probably not binding for most countries, the actual changes are likely to be smaller. 116 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. control of the US authorities. A similar tricky question is how to account for completely new private currencies that do not use any existing national currency as unit o f account. 6.1.2.3 Limited influence As long as the Internet economy stays small compared to the rest of the world economy any effect will be very limited. Countries with a small amount of currency outstanding relative to the size of demand deposits will face much smaller disturbances than the big world currencies. Table 6.3: Currency in circulation in various countries1 9 8 In per cent Currency/GDP Deposits/GDP Currency/central bank liabilities Austria 6 10 43 France 3 19 38 Germany 7 16 63 Italy 6 31 28 Japan 9 24 85 Switzerland 8 18 43 United States 5 45 84 Besides the banks’ willingness to expand deposit money issuance finds a natural upper limit by the contractual obligation to convert this liability on demand back into central bank money. Otherwise at least in the beginning people will not accept digital money. Then, since central bank money can be made scare by reserve requirements the central bank can indirectly influence digital money. As long as the users of digital money instantly redeem it for conventional money it will simply be a new version of existing 1 9 8 1994 data; Source: [BIS 1996] p.5 117 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. payment systems. But in the long term the idea that nobody will pay face-to-face making a separation between conventional and digital money unnecessary is not very realistic. There is yet another brake on an explosive Ml: Since people fear the danger of accidental loss they will be reluctant to store large amounts of value on their hard disk or card. Money, therefore, will not serve as a store of value for the time being. At the same time, people will not use digital money for big transactions if they do not want to be surveyed. Maybe even the issuer limits the maximum amount held per person to reduce his risk exposure. When reserve requirements do not bind, central bank policy relies mainly on signaling. The flow of digital money outside the central bank system might have no real effect1 9 9 since the central bank still influences its users’ believes. Further it seems possible that part of the demand for digital cash is additional with an only weak negative relationship to households’ conventional currency holdings. Reasons are the incompatibility of the two systems and the differences among the products that can be acquired in network and real markets. Altogether for the foreseeable time the aggregate amount of credit supplied by digital money will be rather small.2 0 0 As the impact on the quantity of currencies itself seems to be insignificant, we have to look at the velocity to find the potential cause for big changes. 1 9 9 [Ely 1996] p.3 2 0 0 See also chapter 3.3. 118 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. It has be argued2 0 1 that a more or less complete elimination of central bank issued cash can make the money base rule more effective: If only digital money is held, the currency to deposit ratio c for central bank money is zero. The money multiplier therefore 1 + c 1 simplifies to m = ------- = —, where b denotes the reserve to deposit ratio of the bank. As b + c b the money stock M equals the money multiplier m times the monetary base B, the disappearance of the hardly controllable c makes M easier to direct. A substitution of electronic money for reservable deposits and a reduction of the banking system’s demand for settlement balances take central tools for the control of monetary indicators away from the central bank. We can expect that many new payment products will evolve in the future that require similar adjustments and new policy techniques. Digital money is just one of them. 6.1.3 Velocity of money Irving Fisher’s “equation of exchange” defines income velocity as Y • P GNP v = ------ = -------- . From the introduction of checking accounts over the evolvement of M M electronic funds transfers to electronic and 24-hour banking many monetary innovations increased the income velocity of central bank money. The base velocity has been increasing at a compounded annual rate of 1.5% since 1959. In total the base velocity since 1959 grew by 66 percent,2 0 2 caused by a 300 percent growth in the deposit 2 0 1 [Selgin 1996] 2 0 2 Corrected by foreign currency holdings the base velocity increase by even 200%! 119 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. component. At the same time the currency part of velocity stayed about constant over this period.2 0 3 Digital money as economizer will be only another step in this process, and not dramatic compared to the evolvement of money transfer systems in the past. “Any higher real velocity need not necessarily be inflationary. ... The complexity of financial markets makes profitable, an often requires, numbers of additional payment steps.”204 Digital money, however, could increase the velocity in an unprecedented manner. Depending on its regulation the substitution of national currency could blow income velocity up to infinity. This shrinks the central banks’ currency and deposit holdings further. The domestic demand for central bank money could drop close to zero, if currency and reserve demands vanish. The only lower barrier is the legal tender property, which requires minimum holdings to satisfy tax liabilities. It is questionable, however, if this velocity measure has any kind of economic relevance, because it is observed only in the small remaining government money market that would consist mainly of currency held abroad. The digital money market would be ignored. Reserve requirements and capital ratios can reduce the velocity effect of digital money of course by increasing the demand for government currency. Additionally, in the transition period of changing velocity the narrow monetary aggregates may be useless as targets or indicators according to Py = Mv. 2 0 3 [Jordan 1996] 2 0 4 [Solomon 1991] p.36 120 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The increased intraday turnover suggests that we should see velocity more as a continuous than an end-of-day measure. The discrete stock concept of income velocity could be substituted by the flow quantity of transaction velocity. This measures the rate of use of all “M” money for transactions in all markets per time period. Since it will include not only final GNP transactions but all transactions this flow velocity will be much higher than the current stock measures.2 0 5 Like in any dynamic system a high velocity makes maintaining financial stability more difficult. If the central banks’ instruments with which it controls the money supply do not get more precise at the same time, oversteering might become more common than desirable. 6.2 Seigniorage Governments lose two possibilities to balance their budget: They can no longer count on seigniorage, the difference between the face value of coins and the cost of manufacturing them. They not only have to cede the profits from new issues to the private institutions, but also have to buy back outstanding currency. More dramatically, they have to give up the interest income on the reserve requirements and the currency held instead of government bonds. For the foreseeable future, digital cash and smart cards target only small value transactions. This limits their current substitution potential to the smaller denominations 2 0 5 [Solomon 1991] p.29ff 2 0 6 In 1994 the Federal Reserve remitted about 20 billion US$ to the Treasury. 121 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. in circulation. All currency smaller than or equal to ten US$ amount to about half of all pieces outstanding, but that accounts for only US$ 50 billion corresponding to 13 percent of all circulating currency value.2 0 7 At a long-term annual interest rate of seven percent the yearly interest income loss is bounded by an upper limit of about US$ 3.5 billion. If we assume as in chapter three that every adult US citizen carries a stored value card with a US$ 100 balance, the interest on the outstanding amount of about US$ 20 billion sums up to less than two billion US$ annually. Since the circulating currency grows each year a partial substitution by digital cash will probably merely reduce the growth rate, but not reduce the outstanding amounts. The potential seigniorage loss therefore amounts to less than 773 million,2 0 8 which were the proceeds in 1994. Therefore, the income risks for the central banks impair at most their independence. 2 0 7 [Wenninger 1995] p.5 2 0 8 [Treasury 19961 ] ch.3 122 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 6.4: Seigniorage loss2 0 9 As percentage of the GDP Seigniorage Percentage decline in seigniorage before break even point of central bank is reached Maximum seigniorage reduction Canada 0.31 91 0.15 France 0.28 54 0.08 Germany 0.52 86 0.06 Italy 0.65 91 0.09 Japan 0.42 85 0.06 Switzerland 0.45 88 0.05 United Kingdom 0.28 89 0.14 United States 0.43 93 0.14 An indirect effect could add to the lack o f seigniorage income for the government: If the central bank reduces its holding of public debt by open market operations to offset the purchasing power of the digital monies, a major demander for government bonds leaves the market. The market clearing interest rate for government bonds will raise and the Treasury will have to pay a higher interest rate to finance its debt. 6.3 Control 6.3.1 Shrinking central bank balance sheet More serious is the effect of the shrinking central bank balance sheet and the diminished monetary base. Cash is a large component of the central banks’ liabilities 2 0 9 [BIS 1996] p.8 2 1 0 The maximum of three scenarios: a) if digital money eliminates all banknotes denominations up to US$ 25, b) if every individual carries US$100 digital money c) if digital money eliminates all cash payments up to US$25. 123 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. especially in the G7 countries,2 1 1 which would disappear if digital money spread extensively. “This effect however is not exclusively linked to the introduction of electronic purses. The progressive substitution of banknotes and coins for cashless money has had the same result and, over time, it can be expected that central banks that wish to control the monetary base will be able to find ways of doing so, for example by introducing or increasing the use of required reserves.”21 2 If central banks do not engage in the new market when their proprietary market is dying, their market power dwindles. The central bank needs lots of these assets to implement its policy on the market, for example to buffer the issue of new Treasury bonds. Most depositories build up daylight overdrafts until they can obtain funding from the central bank by either long term borrowing or selling securities. For normal transactions like this smaller central bank balance sheets still suffice. In extreme situations as they appear in foreign exchange markets, however, the central bank must be able to perform huge reserve absorbing operations in short time to offset for example the effect of large purchases. 6.3.2 More power for financial markets Private monies end the possibility for the central bank to influence the money supply and economy. Digital money shifts even more power to the financial markets. The capital of the private sector exceeds any central bank’s foreign exchange reserves by far. Compared to them also the underlying international trade-related flows look like dwarfs. 2 1 1 See table 6.3. 2,2 [EMI 1994] p.9 124 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. International trade for the total of 1995 amounted to about four trillion US$. At the same 213 time international financial markets transacted about one trillion US$ every day! If even higher private funds are moved in the worldwide networks the effect of a central bank operation on the markets could be hardly noticeable. The crisis in the European Monetary System in 1993 made this situation obvious. 6.3.3 International issues Networks know no border. Central bank policies especially with respect to digital money need global coordination. Otherwise regulations can be avoided easily. A perfect national payment system doesn’t make the situation any better if international systems collapse. To avoid that, international clearing mechanisms should reach speeds comparable to the originating value flows. 6.3.4 Disintermeditation Network money makes direct transfers of purchasing power from private agent to private agent possible. Unlike notes the transferred amount does not have a physical limit in size. This eliminates the need for clearing procedures, which the central banks so far used as a point of data collection. If a broad range of unregulated institutions outside the banking system issued decentrally the new digital monies, keeping and enforcing common standards becomes more difficult for the central bank than coping with a limited number of banks. 2 1 3 [Crede 1996] p.3 125 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. However, the disintermediation concerns of banks are only half o f the story. Digital money requires trusted third parties and a number of agents and brokers to work efficiently. In the foreseeable future digital monies are likely to be linked in some way to a respected banking institution to gain access to financial services. 6.3.5 Time lags Private companies have always been reluctant to collect and provide payment information to the central bank. The reason stressed most often is the costs of acquiring the information.21 4 If at all, then information usually is provided with a significant time lag A. When a central bank tries to follow a monetary target M* that includes private monies this information lag can cause severe oscillations in M. To see this assume the central bank follows a simple feedback rule. The central bank reacts to a one time shock that takes M away from its optimal value M* by fm >0 is the derivative of the central banks response function with respect to money deviation. The lag A represents the time between the observation o f M and the policy showing effect in the economy. The solutions2 1 5 of this differential equation are of the form M{t) = M *+(M 0 -M*)eM + ,a > , where p+ico is a complex constant dependent on fr a and A. 2 1 4 See [Meyer 1986] p.26. 2 1 5 [Solomon 19912 ] p.93 126 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Only A=0 ensures a solution free o f oscillations (co=0). For A>0 this feedback rule not only necessarily implies oscillations, but also does not guarantee convergence back to 7 1 M*. The critical value above which the steady state solution is not stable is f m = — . 2A A high A makes it necessary for the central bank to choose a very small fm to avoid a resonance. This means that it cannot follow a tight monetary policy any more. By requiring M to move in a wider band the time lag A actually limits the central bank’s policy to a signaling of the wished direction. If central banks do not want to lose even more control over the money aggregates, they will have to ensure that the average reporting time lags stay at their current levels. This means that also private institutions that engage in issuing money must report their balances like banks instantly. This should pose no major problem: Even in unaccounted systems the outstanding funds are recorded. Only what happens with the money after issuance is a private matter, like in today’s money. Alternatively, central banks could choose monetary targets without an observation lag, for example the daily observable interest rates, to avoid oscillations. 6.4 Confidence Conventional currency is legal tender and backed by holding of US government securities. Private digital money cannot be legal tender and is only backed by a promise of the issuer to honor its value. The confidence in digital money will therefore be lower than in conventional cash. The situation can be improved if digital money balances get for example FDIC guarantees, which makes them ultimately backed by the United States government. 127 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Ensuring a smooth exchange between “real” and “virtual” money over some time by guarantees or adequate regulation can make citizens indifferent among the two kinds of money. This would reduce the danger of importing instability into the national currency. Certifying only selected trustworthy institutions to issue money according to central bank rules produces a similar effect. 6.5 Banking supervision Traditionally, the public expected government to:2 1 6 • enforce basic rales, consumer rights and fair competition • provide law enforcement tools and techniques • assure the integrity, safety and efficiency of the payment system • manage the money supply • issue legal tender As opposed to central bank money, in a private money world the market mechanism determines the issuance of legal tender and the management o f the money supply. If the central bank takes a more active role and bases digital money on its base money or issues it on its own all classic working fields stay the same connected with new challenges in the field of integrity and law enforcement. As new task the central authority has to supervise the issuing banks and regulate their activity in the light of public interest. The issues include merger control and non-bank activities, privacy and consumer rights. 2 1 6 See [Gilbert 1996] 128 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. This can be archived by regulation or technical solutions. Additionally, its support will be needed in case of a crisis to manage disruptions in the operation of the payment systems. 129 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 7 Policy options 7.1 Freely competitive digital monies 7.1.1 Laissez-faire The central banks’ task as a money manager is to balance the flow of money and credit with the needs of the economy. If the new payment systems do not fundamentally disturb this balance, there is no reason to interfere. Under the existing regulation in most countries the banknote monopoly does not extend to digital money.2 1 7 Doing just nothing, the policy of choice of central banks today, makes private digital money possible and encourages system innovation. The market can develop as number and kind of issuers is unrestricted. However, systemic risks may develop and lead to unwanted crises. 7.1.2 From no intervention to free banking 7.1.2.1 Free banking During the free banking period in the United States 1837-1863 state chartered banks issued paper money, which was redeemable in gold and silver. Same state laws restricted currency issues to a fraction of the bank’s capital and imposed penalties on banks that failed to redeem. Although this prevented most banks from overissuing, some occurrences made people suspicious of unfamiliar banknotes. Some banks were intentionally located in remote areas to evade the redemption of their money in gold and 2 1 7 For EU countries see [EMU 1996] p.683 130 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. make overissuance easier.2 1 8 Money traders bought these currencies at significant discounts from the face value depending on the difficulty of returning them to the issuer. Over time clearinghouses developed that provided clearing and settlement of payments, cooperation and mutual help in times of financial crises. The motivation was that a ran on one bank could trigger runs on other banks and destroy the confidence in all private monies.2 1 9 Clearinghouses limited the effect of runs but notably even the best managed banks tended to discontinue redemption in gold during a crisis. The ability to issue currency led to an explosion in the number of banks between 1837 and 1860. At some time more than 10000 different banknotes were in circulation. Accepting money required information on the current value of this currency. Accordingly, money traders published lists with quotes for the circulating currencies. Although this information was widely available and most people used money from known banks, they frequently misestimated the purchasing power of money accepted in trade. The National Banking Act of 1864, which required backing of the currencies with US bonds, and a ten- percent tax on state bank notes introduced in 1865 successfully paved the way back to a single currency. The circulation became unattractive and state banknotes soon disappeared. Since 1935 only the government has the right to issue notes.2 2 1 2 1 8 See [England 1996] and [Schreft 1997] p.71 for details. 2 1 9 [Gilbert 1996] 2 2 0 [Fed 1997] 2 2 1 [Roberts 1997] p.36f 131 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. More optimistic make the experiences in Scotland 1716-1844.2 2 2 Starting in 1727 the notes of the Old “central” Bank established in 1707 by the Scottish parliament had to compete with the Royal Bank that was supported by the parliament in London. After some time of destructive rivalry they switched to peaceful currency competition. Soon also other banks entered the money issuing business. In 1771 all Scottish banks agreed to accept each other’s notes at par which disciplined issuers by mutual control. Not only led this competition to a series of innovations, its stability also made it the currency of choice in northern England. Notably the bankers reserved themselves the right to redeem notes not immediately upon demand, but later with interest to ensure their liquidity. Most free banking areas “preceded the introduction of legal restrictions preventing the currency competition of privately issued bank notes.”2 2 3 Nevertheless, overall experience shows that competition can create quality currencies as soon as reasonable regulations are implemented.224 7.1.2.2 Competitive currencies today? The United States Constitution states: “No state shall ... coin money [nor] emit Bills of Credit [i.e. paper currency]....”2 2 5 2 2 2 [Lynch 1996] p.l05f 2 2 3 [Marmion 1997] p.3 2 2 4 See [Schreft 1997] p.74. 2 2 5 United States Constitution - Article I - Section. 10. - Clause 1 132 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. This clause does not mention private money. Regulations, however, prevent private companies from issuing plain currency:2 2 6 Firstly, the prohibitive ten percent tax on state banknotes and bond collateral requirements of the National Banking Act makes it financially unattractive and secondly, private money cannot be legal tender227 which subjects it to competition. Thirdly, 18 U.S.C. 336 provides “Whoever makes, issues, circulates, or pays out any note, check, memorandum, token or other obligation for a sum of less than $1, intended to circulate as money or to be received or used in lieu o f lawful money of the United States” is subject to a fine and imprisonment. There is simply no demand for private money that can be issued under these circumstances. A main obstacle of any private money is that it has to be accepted by other agents. It is exactly this network externality that banished company scrip to closed systems. Only a few money deviates like travelers’ checks and demand deposits were able to establish themselves as new payment instruments. And as long as banks and settlement institutions do not accept private money, the public is unlikely to hold it. Today, technology offers new possibilities to issue private currency. The Internet eases the mass issuance and circulation by providing a cheap currency transfer infrastructure2 2 8 and information exchange. Real time conversion between units makes strong competition possible. Finally, the legal vacuum surrounding digital money and its design allows avoidance and circumvention of the existing regulation. 2 2 6 See [Treasury 19961 ] appendix 3. 2 2 7 See 31 U.S.C. 5103. 133 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 7.1.2.3 Private currency in the economic literature One of the most emphatic proponents of privately issued money was Friedrich von Hayek. In his book “The denationalization of money” he states that money does not have to be created as legal tender.2 2 9 Although money is not a natural monopoly, governments usually suppress all other money. “The past instability of the market economy is the consequence of the exclusion of the most important regulator of the market mechanism, money, from itself being regulated by the market process.”2 3 0 Possibly, competing currencies comes closer to the characteristics the user desires than monopolistic currencies. If required backing reserves and links to the established banking system of digital money are kept small, we have a kind of “free” money as envisioned by Hayek. • 231 The argument that regulation is unnecessary bases on two mam assumptions: • Self-regulation: Market forces discipline issuers. • Independence: The confidence in issuers is independent. Market preferences should force unregulated institutions to hold enough liquid balances to satisfy redemption demand at all times. High efficiency and low cost can make this private money system Pareto-optimal for society. Because the issuers have to arrange lines of credit with banks to meet their commitments they are disciplined by the 2 2 8 [Mantonis 1995] p.2 2 2 9 [Hayek 1976] p.30 2 3 0 [Hayek 1976] p.79 134 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. regulated banks’ conditions and therefore indirectly by the policy of the central bank. This should induce issuers to invest only in high quality assets to maintain credit worthiness and confidence.23 2 Hayek notes “money is the one thing competition would not make cheap, because its attractiveness rests on it preserving its dearness.”2 3 3 However, once people hold a particular currency severe time inconsistencies appear in a completely free banking environment. Since the issuers have a default option, they can decide to reduce their outstanding liabilities by inflating their currency. ‘I f suppliers of currency cannot commit on their future actions then competition may loose its bite.”234 The banking panics in the 19th and 20th century prove the inability of markets to regulate themselves. The second assumption of independent confidence is even stronger. Of course competing monies must be distinctive, because only then the reputation effects can restrict overissuance.2 3 5 If the monies were not distinguishable each issuer would have an incentive to issue more on the cost of the other issuers, eventually driving money out of existence. But even with distinctive monies this assumption requires that consumers 2 3 1 See i.e. [Hayek 1976] p.43. 2 3 2 See i.e. [Gilbert 1996], 2 3 3 [Hayek 1976] p.74 2 3 4 [Marimon 1997] p.5 2 3 5 See [Hayek 1976] p.38. 2 3 6 [Schreft 1997] p.67 135 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ignore the effect of instability among other issuers on the whole payment system and ultimately on their own currency. 7.1.3 Private suppliers of digital money 7.1.3.1 Financial institutions Financial institutions are regulated to produce public confidence in their payment instruments and systems. They are supervised and regularly examined by federal agencies, which check the asset quality and risk management skills of the company. Possible asset holdings and operations of these companies are restricted, but are at the same time protected by the federal deposit insurance. Regulated banks play an important role in the payment system: They ensure a minimum level of security and set standards, they protect against fraud and act as clearinghouse. 7.1.3.2 Non-financial institutions Non-financial institutions like telephone or mass transit companies are free from the restrictions their regulated sister institutions have to face. This gives them a competitive advantage. Although banks and non-banks might perform similar functions, their profit margins will differ due to unfair regulation. This explains why all major digital money systems have been developed and are owned by non-banks. If non- financial institutions expand their already existent single purpose payment systems or introduce new ones, they could sooner or later dominate the money supply. Successful alternatives to clearing over a bank marginalize the established commercial and central banks’ influence. 136 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Of course private payment systems could become more secure than government money, as it is already today observable in the bond market. However, the public confidence in unregulated and uninsured issuers will be far from stable. One of the main arguments for a regulated banking system was exactly consumer protection by minimizing the risk of failure and building up confidence. Additionally, the abundance of non-financial institutions is difficult to supervise. And because they are not periodically examined, they offer no easy accessible audit trail for law enforcement. Not surprisingly, the European Monetary Institute23 7 and the United States Treasury2 3 8 are discussing ways to restrict the issuance of digital money. 7.1.3.3 Financial innovation Financial innovation can evolve in isolation from the established banking sector. Grigg2 3 9 looks at the history of US dollar offshore intermediation started by S.G. Warburg in 1955. From the very beginning this was doomed to be a risky, slow growing enterprise in an unregulated market without any access to central bank facilities. Initially very cautious and growing very slowly, Warburg kept high reserves and managed several crises. Finally, the acceptance by the Federal Reserve gave the ignition to fractional banking, since now the lending guarantee by the central bank made the risks manageable. 2 3 7 [EMI 1994] 2 3 8 [Treasury 19961 ] 2 3 9 [Grigg 1996] p.6 137 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. “The interbank liabilities trade wasn’t enough to make them feeling save enough. The credibility of the banks assets was insufficient.”24 0 This development happens again in Internet banking. Small non-financial companies outside the banking system introduce new products. In case of success they will be either integrated in the regulative environment of banking or acquired by banks that are subject to central banks’ influence. In both cases they gain access to emergency lending and can therefore take up fractional banking. 7.1.4 A free private money market 7.1.4.1 Private issuers’ working areas a) Cash flow management The issuer has only to manage distribution and redemption of his currency and oversee its regional distribution; all other functions like the management of the backing funds can conveniently be outsourced. Mantonis2 4 1 emphasizes the importance of a sensible cash flow administration. Because money can be returned almost instantly, and positive feedback effects are likely to be even stronger with private monies, changes in private monetary policy and public relations must be managed very diligently. Tracking the redemption patterns might help to ensure the necessary liquidity at all times. In order to avoid self-fulfilling prophecies 2 4 0 [Grigg 1996] p.6 2 4 1 [Mantonis 1995] p.7 138 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. the issuer must convince speculators that he will be able to support the value of the currency at all times. Only this can smooth the traders’ expectations about the future. Notably, the issuers do not have to guarantee instant redemption upon demand. It suffices to offer credible redemption at a specific day in the future. However, they should still care about their money’s value development to keep confidence high. They can do so by actively buying and selling currency similar to the open-market operations of a central bank. b) Backing management First and foremost, money serves as medium of exchange. The issuer, however, wants to build up a good reputation and high added value to introduce his currency as a store of value. This keeps the money in circulation for a longer time. The consumer will have to make an investment decision when he chooses his money. The question of backing - credit systems versus full backing - and its quality - Treasuries or high-risk venture capital - will become the basis of competition among issuers. Unfortunately, offering high interest usually brings a riskier backing policy with it and therefore stands in open conflict with a good reputation. In order to avoid inability to redeem the currency, the issuer should choose his main backing asset close to the one in which he guarantees redemption. In this case redemption would never be endangered, independent of the market situation. No matter if a stock or bond portfolio backs it, a digital money unit can have some similarity with a 139 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. share in a mutual fund,24 2 which is constantly marked to market. With no advantage of redeeming money first, the motivation for runs disappears. Competition for the consumers’ favor requires a professional management of the backing funds similar to investment funds. c) Rating performance Private monies will be subject to rating by special rating agencies. These ratings solve the asymmetric information problem: They indicate the financial integrity of the issuer and determine the market value of each currency. Establishing a good currency reputation can produce a well-known brand name as a side effect. Especially financial institutions can profit from such a link to a respected currency. Dependent on their rating currencies will not trade at par value to the US dollar but at a discount or agio. People will no longer have a uniform currency but will have to inform themselves about the value of the currency in advance of every transaction.2 4 3 d) Competition Induced by competition, private money may have additional properties preferred by the household. First and foremost, basic attributes of money - acceptability, stability, low risk and certainty of payment - have to be ensured. Government as monopoly supplier did not have any incentive to give money features on top of this basic design. 2 4 2 For fund managing companies issuing money backed with their funds could be just a new way to market their products to a broader public. A company failing to deliver the guaranteed backing in this case commits fraud. In contrast to government money this default will have a sequel in court. 2 4 3 [Schreft 1997] p.65 140 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. People might favor money that pays interest among the various providers. Now they can freely choose the best money, money that is not only widely accepted but also durable and usable as a store of value. Success as money issuer requires developing the same attributes we can observe in every competitive environment. Low fees and value added services build up a loyal customer base. By establishing long-term trust in his currency, and linking this to his brand name he might be able to create a profitable niche. If people believe in a currency each note will be longer in circulation and increase the issuer’s interest income. Every supplier o f money in this competitive environment is able to manipulate his own distinct prices. If other currencies suffer from instability, well-managed private money units will appreciate relative to the other and increase their market share. However, this might also undermine the confidence in the whole private money system. When government and several private monies coexist there is always an unrealized profit potential since inflation in rival currencies offers the chance to gain additional market share. Money issuers can also hope that countries with monetary instability will use their money without expecting any added value.244 This would of course increase the revenues, but I doubt that countries with monetary instability have the network infrastructure today to use digital money efficiently. 2 4 4 [Mantonis 1995] p.7 141 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 7.1.4.2 Clearing and settlement Exchanges between multiple private monies require a clearing system. But even if there is only one digital currency involved, the bank from which funds are withdrawn does not have to belong to the bank where the merchant has his account. Depositing money directly to your account at another bank without having an account there seems to be impossible today.245 But digital money will require exactly this: a deposit clearing system that guarantees final settlement of the banks net position between issued and redeemed money. The clearinghouse necessity shows the difference between old and new money very clearly: Whereas trades in one conventional money are final immediately, positions in several digital monies are not final - they have to be setded against each other - possibly in conventional money. A private clearinghouse cannot guarantee settlement finality, since the default of a few big borrowers could bring it into default, too. This would underline the credit character of digital money. People think, however, that the central bank will ultimately help the clearing system over liquidity problems: ’’Payments over many private networks in effect constitute money because they are perceived as such.”2 4 6 A presumably privately organized clearinghouse is the more elegant way to holding reserves at each and every other bank. It must be more open than today’s 2 4 5 [Philips 1996] 2 4 6 [Roberds 1993] p.4f 142 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. settlement systems to be also accessible for providers originating in the non-bank sector. Today this is not generally the case. The European Monetary institute states that “as a rule, only central banks and credit institutions ... can be admitted as direct participants in funds transfer systems which process third party payments.”2 4 7 As exception it only sees institutions that have little risk of failure or are supervised by a recognized authority. But especially small institutions rely on the services of such a clearinghouse to avoid the risk of illiquidity caused by delays. The settlement of net amounts has to occur at least daily but preferably instantaneously to avoid instabilities.2 4 8 Since losses from the default of a member have to be bome by the clearinghouse, it will set minimum financial standards for its members and frequently control them. Similar to a central bank it would provide a mechanism of clearing and redemption of digital balances and force issuers to keep a sound balance between digital balances and reserve backing. As we see in the standardized futures market, clearinghouses usually require some form of security backing called margin.2 4 9 This kind of free-market mechanism could force issuers to keep some minimum backing to tackle the settlement risk. As a side product this would also put an upper limit on overissuance and ensure a minimum degree of creditworthiness. The existence of clearinghouse guarantees - if they are credible - enhances consumer confidence in private money. 2 4 7 [EMI 1997] p.31 2 4 8 See chapter 4.5. 2 4 9 Whereas the Federal Reserve determines the margins for buying stock on credit (50%), the clearinghouses set margins in the derivative markets themselves. 143 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 7.1.4.3 Information implications Accepting private money requires information. This information in likely to be asymmetrically distributed among customers and issuers. If people do not evaluate all information then they might see different monies as substitutes and adverse selection will drive the better monies from the market.2 5 0 Kocherlakota2 5 1 studies difficult trading environments with limits on outside money holdings in which people cannot commit about future actions. He excludes the possibility of barter by allowing one trading partner at maximum to have the good the counterpart desires. Only two kind of transactions exist: Monetary transactions, which combine production with the acquisition of money, and non-monetary transactions, which imply production based on good faith, for example some form of credit like digital money. If the consumer happens to have outside money and the producer has not, they can choose whether they want to settle the transaction in outside or inside money. In all other cases, especially if their money holdings are equal, one side can produce in exchange for inside money in the hope that it will get something back it in a mirrored situation in the future. It will only do that if its information about the trading partner does not show defaults in the past. If this important attribute is unobservable, bad money will prevail by adverse selection. 2 5 0 See [Schreft 1997] p.68 but also [Hayek 1976] p.34. 2 5 1 [Kocherlakota 1996] 2 5 2 Inside money in this model can be anything from credit to private digital money. 144 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Accepting inside money always includes the risk of default. As soon as an issuer has digital money outstanding he additionally faces a moral hazard. He can realize an advantage on the cost of his customers by defaulting without his behavior being observed. The history in this economy consists of two parts: outside money holdings and public records of all past actions that are updated with a random lag. Money proofs the entitlement to gain in the future because of past production. But without money the knowledge about past actions is incomplete. The model shows the advantage of up-to- date information systems and how lags affect the optimum. Specifically, information time lags influence the mix of monetary and non monetary transactions. An infinite information time lag makes private information out of the (credit) history of each issuer. The default temptation makes digital money impossible. Only outside monies can be used in exchange. In the opposite extreme, ad hoc information means that each issuer’s actions are public information. Then, outside money is inessential.2 5 3 Kocherlakota shows that a society with more regular updating of the participants behavior, for example the redemption policy of a money issuer, is at least as well off as the less frequently updating economy. The higher detection risk and the shorter time span of enjoying undetected the advantages of defaulting reduce the issuer’s gain of defaulting.2 5 4 2 5 3 [Kocherlakota] p.2 2 5 4 [Kocherlakota] p. 14 145 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. If the discount factor in this economy is sufficiently close to one the optimal behavior for the household is • To produce only in exchange for outside money, if the time lag in infinite • Independent of getting or not getting outside money if the time lag is zero • Dependent on outside money transactions if the time lag is big. The utility of the household is strictly decreasing with the time lag. In case o f an infinite time lag the household make the optimal choice to use only outside money. In order to reach a higher utility he has to use the additional payment possibility to some extent as soon as there is some information available.25 5 For a sufficiently high discount factor also a society with an information time lag close to but not equal to zero runs at the optimum independent of the distribution of outside money. If credit or other private digital money is trustable, outside money is irrelevant. Kocherlakota concludes, “In general this setting makes it desirable to use both (outside) money and some form of credit to accomplish trade.”2 5 6 The rapid and correct information, that electronic networks provide, is important to make inside money a widely used alternative to outside and secure central money, if there is no commitment for private issuers. An alternative to accelerated and accurate information flows is the use of some technology to make commitment possible, for example laws. 2 5 5 [Kocherlakota 1996] p. 17 2 5 6 [Kocherlakota 1996] p.18 146 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 7.1.5 A central bank without power? Even if the central bank has no regulatory power in a free market scenario, it will still play an important rule:2 5 7 As an institution outside the competitive field it has a good chance to maintain its rule as the clearinghouse of clearinghouses that provides settlement finality. Its ability to offer ultimate clearing comes from its foundation as institution that serves the common interest. By issuing legal tender it is the only source of payment tokens that can be used for the settlement of tax obligations. Finally, it could maintain its position as one big source of intraday credit to meet account balance requirements. The clearing and settlement functions require of course sufficient assets, which now have to come from some other source than issuing currency. If its position as market leader stays unchallenged it can still determine the terms at which institutions can lend excess funds or borrow for debt accumulated during the day. Although central bank money is not held overnight, is will still serve as medium of exchange and unit of account during the day. Maintaining its purchasing power therefore still is at the focus of the central banks’ activity. 2 5 7 [Jordan 1996] p.5 147 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 7.2 Supervision of competing monies 7.2.1 Necessity of regulation Since Adam Smith many economists have stressed the importance of government regulation in maintaining a currency’s quality.2 5 8 Information - especially in the case of small money holdings - is a public good, most efficiently supplied by only one institution. Free riders will make sure that it will be undersupplied without special regulation. Government intervention is also required in case of a mispricement of risk: If the systemic risk borne by others is not considered by the payment system users, the choice of their payments portfolio will be inefficient. As mentioned above users could for example rely on the size of the payment system, whose failure will be stopped by government help. In making their decisions they undervalue the total disutility of the risks connected to their payment choice. By unjustly preferring issuers that are “too big to fail” and ignoring the govemment-bome cost of risk, their choice will not be socially optimal. In figure 7.1 the subjective indifference curve leads to a payment system choice that is clearly suboptimal judged by the objective indifference curve. 2 5 8 See [Schreft 1997] p.69 for a short historical overview. 148 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Risk worse outcome efficient frontier subjective objective indifference indifference Cost curve curves Figure 7.1: Private underestimation of risk2 5 9 The government has therefore the duty to price these risks or to regulate the choices toward the use of less risky, but probably more costly systems. The European Monetary Institute states that unregulated digital monies are “not compatible with fundamental central bank responsibilities for maintaining the integrity, stability and efficiency of its county’s payments system and for the conduct of monetary policy.”2 6 0 Issuing digital money provides the bank with conventional money, which it can lend or invest. It can make profit at zero cost. Logically, the company will be tempted to issue more. Without any regulation we get an electronic Ponzi scheme that would finally degenerate in a commodity system with no money at all. Regulation o f private issuers 2 5 9 [Berger 1996] p.703f 2 6 0 [EMI 1994] p. 1 1 149 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. reduces their possible float collection and interest free loan opportunities. Especially if few payment systems dominate the market, the importance of efficient and cheap payments for public welfare suggests the use of antitrust measures, hi order to provide a working payment system the central bank can rely of the following measures: 7.2.2 Technical restrictions As supervisor of payment systems, the central bank can impose technical restrictions on money schemes. As extreme solution it could require all systems to be fully auditable. Alternatively, in order to reduce the crime potential in offline cash systems it could for example limit the time a value token is valid, after which time it would either expire worthless or loose its ability to circulate. This would require network wide synchronized time stamps. Similarly, a limit on the amount of stored tokens or daily transactions per person or card would make electronic money unattractive for illegal activities. Besides legal pressure the government can support certain developments by its central position in payment services and settlement. There is reason to hope that the risk exposure will induce the market to introduce some of these measures on its own. To prevent developments that could endanger security or function of the payment system, creation and administration of electronic payments should be supervised. Besides fighting technical and counterfeiting dangers, protecting privacy by keeping payment information separate from commercial data collection should also be of concern for the central bank. 2 6 1 [Cuadras 1997] 150 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 7.2.3 Banking regulation and supervision The softest way of regulation would be to require non-depository digital money issuers to conduct their payments through a regulated bank. Stronger supervision could imply that only certain institutions that satisfy the central bank’s minimum requirements are certified to issue money. These institutions would issue their money according to central bank rules and give it on to other interested institutions, which could use it as a base for their own currency. Whereas the monies under central bank supervision could be designed to be FDIC insured, those other currencies would then not enjoy this protection, but perhaps be able to offer more favorable conditions. If the FDIC wants to stick to its current analysis,2 6 2 digital money for small transactions may be excepted from this protection to keep the convenience of the money attributes. All in all, banking supervision can limit the systemic risk in financial markets. Unanimously the bank of international settlements (BIS)2 6 3 and the European Monetary Institute (EMI) find that since electronic money requires deposit taking it would be natural to apply the existing standards for checkable bank deposits.26 4 “The reasons which lead public authority to reserve deposit taking to a specific category of 2 6 2 Digital money issued by banks will be insured if the underlying funds remain in a customer’s account until the value is transferred to a merchant Funds placed in a reserve or general liability account held by the issuing bank to pay merchants, however, are not insured. (FDIC General Counsel’s Opinion No.8, 61 Fed. Reg. 40490, August 1996 in [Treasury 19961 ] appendix I.) 2 6 3 [BIS 1996] p.9 2 6 4 Compare [Hayek 1976] p.41. 151 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. institutions should also apply to the issuers [of digital money].”265 For example, if a bank wants to issue money backed with mutual funds, the central bank could treat this payment instrument similar to a money market mutual fund with check writing privileges.2 6 6 Restricting the issuance of digital money to banks has a number of advantages: Infrastructures for clearing, settlement and information exchange between central bank and financial intermediaries already exist and all market participants are subject to the same regulatory framework. Not surprisingly therefore, a European Working Group on payment systems proposes to restrict the issuance of electronic money to financial institutions that are subject to banking supervision.267 “The purchasing power loaded in a multi-purpose prepaid card represents, for the issuer, a source of funds equivalent in its economic effect to deposit-taking. Therefore the right to issue electronic purses needs to be restricted to credit institutions in order to: 1) protect the integrity of the retail payment system; 2) protect consumers against the consequence of the failure of the issuers; 3) facilitate the conduct of monetary policy; and 4) ensure fair competition between issuing institutions.”2 6 8 2 6 5 [EMI 1994] p.8 2 6 6 [Good 1997] p.23 2 6 7 [BIS 1996] p.7 2 6 8 [EMI 1994] p.2 152 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Although this point of view has not been adopted in the United States26 9 so far, it would make sure that the existing legal framework could be used: Digital money would be subject to regulatory oversight and backed by appropriate assets. The United States Treasury suggests that if the “issuance of electronic cash involves the receipt of deposits”270 federal laws2 7 1 that restrict deposit taking by non-banks could be applied to digital money. “For example, federal laws limit the amount of ‘loans’ that banks may make to a single borrower and, thereby, protect banks form excessive concentrations of credit risk. ... Banks holding electronic cash issued by third parties (banks or non-banks) may be deemed to have made a ‘loan’ to the issuer in the amount of the electronic cash. The lending limits would, thus, restrict the amount of such third party electronic cash that a bank could hold to 15 per cent of its capital.”2 7 2 That way, all or parts of the existing regimes could be extended to non-financial institutions, for example reserve requirements, minimum capital levels, monitoring, licensing, bonding and deposit insurance.2 7 3 Existing law could limit the types of firms that issue electronic money services. Alternatively, new special regulation for non- 2 6 9 One reason for the US hesitations to apply banking regulations might be that banks in the USA face more restrictions of their permissible activities than in Europe. 2 7 0 [Treasury 19961 ] chapter 3 2 7 1 For example U.S.C. 1831t and Section 21 of Glass Steagall Act 2 7 2 [Treasury 19961 ] appendix 4 2 7 3 For example, American Express has to back all of its checks’ outstanding value with liquid instruments such as cash or low-risk bonds. See [Radigan 1997]. 153 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. financial institutions could be tailored to their needs, but this will only further complicate the legal code. In any case the central bank has to keep the balance between regulation and innovation. The tradeoff between fair treatment of the banking sector and risk minimization on the one hand and healthy competition on the other require a cautious timing to avoid unwanted results. 7.2.4 Clearing In order to limit the systemic risk, reducing the interdependence among banks must have a priority. Lacking an intermediator the default of one market participant requires recalculating the financial position of all others, which most likely causes additional defaults. When a clearinghouse is the counterparty to all trades, it picks up the losses if one party cannot deliver. If the clearinghouse maintains liquidity at all times chained bankruptcies cannot happen. But only a government institution can guarantee finality independent of the size because of its lending power. Each central bank could clear its country’s institutes and include the guarantee that any value transfer is valid and final, independent of the final settlement. 7.2.5 Central bank lending Even if it is not the central bank but the issuer who guarantees the money value, there remains still the necessity to dampen the effects of speculative runs. As lender of last resort the central bank can solve temporary difficulties and avoid the spread of unexpected deposit drains throughout the payment system. Only the central bank has the 154 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. credibility and possibility to lend infinitely over a short time without losing the trust in its value tokens. The history of the free banking period has shown that private money issuers are vulnerable to mns. Private clearinghouses developed because a failure of one bank undermines the confidence in the whole payment system. Therefore mutual help was preferred against the reliance on pure market mechanisms.2 7 4 Since the success of such clearinghouses is highest if most of the money issuers are connected and competition might lead to conflict of interest among its members such an institution should stand outside of the market forces. A central bank could offer the issuers neutral clearing services and access to unlimited central bank borrowing in certain situations. Free access to central bank lending encourages the banks to invest in less liquid assets since the central bank bears the redemption risk. In order to avoid distortions the central banks will want to put regulations as mentioned in 7.2.3 in place to force issuers into a less risky strategy. To accomplish that the central bank needs to have jurisdiction over all payment service companies. Similar to banks today, all issuers that do fractional banking will then be regulated to ensure their stability. 7.2.6 Loss of control Private monies end the possibility for the central bank to influence the money supply and economy. Governments can no longer count on seigniorage as a possibility to balance their budget. Digital money shifts even more power to the financial markets. As mentioned above, the capital of the private sector exceeds any central bank’s foreign 2 7 4 [Gilbert 1996] 155 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. exchange reserves by dimensions. If even more private funds are moved in the worldwide networks the importance of central bank operations shrinks more and more. If the central bank follows a money target all institutions — including non- financial ones - have at least to report outstanding liabilities used for payment. Otherwise no monetary policy will be possible. As far as possible this should include foreign issuers. The collection of these data in unnecessary if the central bank follows only interest rate targets. In this case the central bank does not need any information from the financial system, because it relies on publicly available market data. 7.2.7 Central bank monetary policy? 7.2.7.1 Reserve requirements The creation of digital money can be limited by reserve requirements. If the central bank wanted to make electronic money issues - quite in contrast to the daily credit creation of banks - completely neutral it would have to introduce a 100% reserve requirement T dm on outstanding digital money. These reserve requirements on digital money would also reduce the seigniorage loss for the central government. It is possible to set the reserve tax at a level that exactly offsets the seigniorage loss. But since reserves work like a tax on digital money this would slow down development. When the market will have reached some maturity, however, this will be a feasible option. Once there are reserve requirements, the new payment systems rest on a distinct central bank base money again.27 5 2 7 5 See chapter 7.3.3. 156 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1.2.1.2 Open market operations Avoiding the unwanted side effects of reserve requirements the central bank can also try to manage digital cash by open market operations to absorb excess reserves. If the central bank wants to keep the total purchasing power constant it has to sell one dollar of its assets, for example government bonds, to offset the effect of one digital dollar that replaces conventional cash. If the central bank does not view digital money as currency and therefore as part of M l it would still have to redeem (1-roM) dollar to absorb the additional excess reserves the bank received in exchange for digital cash. The central bank can also use open market operations to sustain infinite central bank lending and to always offer the right amount of reserves. Without any possibility to get approximately true values of the monetary aggregates the central banks could choose just to follow the “feel of the market”, which means just setting the Federal funds rate. “The rule of money aggregates has been downgraded, the role of interest rates and underlying market forces upgraded in an information-based policy shift.”276 Since the convertibility of digital money into “real” value is an important prerequisite to make digital money acceptable as store of value, the central bank can use off-balance-sheet transactions for example in the futures market to stabilize the exchange between “real” and “virtual” money. 2 7 6 [Solomon 1991] p.44 157 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 7.2.8 International coordination 7.2.8.1 International agreements Mutual help among central banks strengthens their market power. A network of central banks must react at least as fast as its private counterparts to make quick mutual extension of credit in any currency possible. Furthermore, current settlement systems like at the bank of international settlements should include real-time finality in international digital money payments. 7.2.8.2 International central bank Tatsao2 7 7 suggests a separate monetary authority for the cyberspace that regulates digital money issues worldwide. This institution would work much like the existing central banks, supervising, providing insurance and acting as a lender of last resort. It could use some established national currency as base money whose central bank agrees and is capable to manage this additional responsibility. However, many countries may not be willing to accept this loss of national sovereignty. 7.3 Centralized monetary systems 7.3.1 Prohibition? If regulation is deemed to be impossible or too difficult and the effects of digital money too weighty, the prohibition of digital money will be the only choice. However, 2 7 7 [Tatsao 1996] p. 11 158 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. such a behavior should be the ultimate measure because it automatically destroys any advantage that digital money may provide. 7.3.2 Central bank as issuer in competition One imaginable reason for a separate government money would be to offer a better alternative to badly managed private monies and therefore forcing them to ensure at least the minimum standards offered by the government money. “In addition, the use of digital money could develop up to a point where the ability of customers to pay with notes and coins could be threatened. This would be in contradiction with the legal tender regulations ....”2 7 8 Nevertheless, a central bank issuing digital money in competition with private institutions is not very reasonable. Because of the per-definition zero credit risk of its money a central bank has an unfair advantage over other issuers. If the markets work in a satisfactory way then there exists not reason why a government institution should be allowed to distort competition. Interference would therefore “be in contradiction with the long term trends that have led central banks to withdraw from competition with the banking sector and to concentrate on the oversight of payment systems and the provision of interbank services.”2 7 9 2 7 8 [EMI 1994] p. 10 2 7 9 [EMI 1994] p. 11 159 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 7.3.3 Central hank as issuer in competition with distinct base money 7.3.3.1 Mechanism It is imaginable that there are two different kinds of money in the economy: A per definition stable central bank money whose redemption is guaranteed by the government but which has no additional features. And several private monies, which promise redemption in government money and pay interest or offer additional services. Those would be the preferred medium of exchange and be regulated only by reserve requirements. Issuers, possibly only banks, can only offer payment systems based on deposited government money. The money creation follows the same laws as the well-known credit money based on demand deposits. These debit-like monies continue to be claims on the fiat central bank money. The current situation changes only with respect to the complete elimination of banknotes. The clearing deposits of financial institutions on the books of the Federal Reserve remain, too. Since banks simply exchange the returned federal banknotes for federal deposit claims finally the entire stock of money will be in that form. 7.3.3.2 Commitment issues Marmion2 8 0 studies an economy with a monopolized money supply. Only the government can manipulate the price level. Government money competes with perfect substitutes like foreign or privately issued digital monies drawn from interest bearing deposits. Since government expenses are assumed to be fixed any loss of seigniorage has 2 8 0 [Marmion 1997] p.21-28 160 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. to be financed by distortive taxes. Marmion finds that if the government acts as a Ramsey planner, which means that it maximizes the utility of its citizens, it will follow the Friedman rule of zero nominal interest rates. Since the private issuers have to cover their expenses in addition, government money dominates private money. Then, because the return on private digital money must be negative, it cannot circulate.2 8 1 In the no commitment case2 8 2 the representative government has an incentive to deviate from the Friedman rule to reduce its liabilities and the use of distortive taxes. In the case of deviation the household following a trigger strategy reduces its holdings of government money to zero. In an economy without other monies this implies punishment for government default by the complete annihilation of cash, making consumption forever impossible and reducing utility to a very low level. The presence of alternative digital monies as substitutes, however, makes the punishment less severe since people can still consume by simply switching to another currency. The possibility to revert to an equilibrium with only digital money bounds the utility loss of government default. This weakens the disciplinary effect of reputations2 8 3 and makes a time inconsistent behavior more likely. Reserve requirements did not matter in the commitment case since circulating digital money was unprofitable anyway. This changes in the non-commitment case: The 2 8 1 [Marmion 1997] p.21f 2 8 2 [Marmion 1997] p.25 2 8 3 Only if we exclude the possibility of an immediate exchange the utility loss caused by a liquidity crunch may be enough to deter government from deviating. 161 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. government only can collect interest income on the digital money reserves if the price level is finite. After a deviation, however, the reserves are valueless. Since the price level is infinite they can be provided at no cost by the issuer, but that also means no income for the central bank. The potential loss of this income helps sustain a commitment solution. Reserve requirements influence the competition between currency and digital money. Independently chosen reserve requirements make it more attractive for the government to stick to the commitment solution because they enhance the disciplinary effect of the households’ switch to digital money. 1 3 3 .3 Regulation of issuers Issuing digital money based on central bank money resembles deposit based credit issuance in every way. The mere difference is that this credit enjoys wide acceptance as medium of exchange. Therefore, the regulatory framework for all digital money issuers must be exactly the same as for banks to avoid distortions. Otherwise, all banks will sooner or later change their legal status and use digital money as a credit substitute. The confidence in digital money could build on required reserves in conventional central bank cash. When the authorization to issue digital money is restricted to banks, all money value could be guaranteed by FDIC insurance. All other properties of digital money raise similar questions as in the free banking scenario. Banking supervision, clearing and settlement and possible regulatory measures raise the same controversies.2 8 4 2 8 4 See chapter 7.2. 162 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. If the central bank has only to cope with regulated banks, its operations - including monetary policy, supervision and statistics - will become no more difficult than they are now. 7.3.3.4 The position of the central bank a) Management of the distribution of digital money When the central bank issues the base money it holds a central position: By its policy instruments the central bank can influence the creation and distribution of digital money. By choice of reserve requirements it determines the amount of digital money flowing through the economy. By selection o f issuing institutions, the control of banking mergers and - if necessary - appropriate incentives it can support the nation-wide accessibility of digital money systems. Because the exchange of digital money in central bank money does not require a foreign exchange the Tobin tax2 8 5 can be used to bring the international currency speculation back to a healthy level. b) Management of digital money’s derivatives When digital money spreads it is just a matter of time until trades based on this new instrument evolve. In order to avoid speculative distortions in the redemption value and to keep the financial system free from runs the central bank can use open market operations. Further it will want to check new types of derivatives with respect to their soundness before they enter the market. 2 8 5 See chapter 4.5.5.3. 163 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. c) Management of the base money The stability of the whole digital money system rests on the soundness of the underlying base money. The responsibility of the central bank to ensure currency stability will therefore be as big as before. As shown in chapter 7.3.3.2 failure to implement a proper policy will end up in a scenario de facto without any base money. 7.3.4 Central bank as exclusive issuer Central banks following this policy do not have to operate the electronic money systems themselves; they solely have to reserve the right of issuing value. The system operators can still offer added value features. If the central banks decide to issue digital money by themselves, essentially nothing will change. All current policy instruments will work the same way as today. The central bank has full control over the outstanding amount of digital money and therefore no measurement problems. The same is true for velocity, but its value will reflect the higher usage frequency of digital money. An exchange between “real” and “virtual” money reduces to a change of payment instruments, the denomination and underlying central bank policy is the same. The Bank of Finland followed this strategy. By setting a market standard before other non-compatible systems could spread it gained the position of controlling the implementation of digital money in Finland. The electronic benefits transfer (EBT) that 164 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. will replace food stamps in the United States by 1999 can be viewed as the American prototype of digital money issued by a government.2 8 6 Unfortunately, a monopoly will slow down the innovative power of the private sector and cut the competition down to national currencies. This competition will nevertheless increase due to the transnationality287 of the Internet. Although there is only one national currency government strategies might still be time inconsistent like in currency competition, if foreign digital monies become more and more accepted in the national economy. 7.4 Critique Generally, governments should only interfere if there is evidence of market failure. Despite the use of advanced technology, however, the introduction of new units of account will unnecessarily lead to confusion and complicate markets. The most important market signals, the prices, will become harder to observe. National currencies are already standing in competition with each other. Since they become more and more accessible and accepted in other countries, every citizen has the option to choose the worldwide best money. Properly managed national money has at least the same long-term credibility as privately issued monies, and states disappear far more seldom than the even biggest corporations.2 8 8 A 2 8 6 See [Solomon 1997] p.79. 2 8 7 See chapter 4.4. 2 8 8 [Bemkopf 1996] p.7 165 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 8 Conclusion Governments have a responsibility to ensure the circulation of high-quality currencies. Despite all the excitement about digital money nobody should forget the lessons of economic history. In the end only the citizens’ welfare counts. Risky experiments with unregulated monies are clearly a not responsible policy. The integrity of the worldwide payment system does not permit a “wait-and- see”2 8 9 strategy. Rather, central banks should provide clear, open guidelines to direct investments towards sound and welfare maximizing systems. ■ 289 [Fed 1996] 166 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 9 Bibliography [Artmann 1997] [Baumol 1952] [Berentsen 1997] [Berger 1996] [Bemkopf 1996] [BIS 1996] [Blanchard 1989] [Brandel 1997] [Campbell 1988] [Congress 1996] [Crede 1996] Artmann, Christian and Strasser, Georg: Recht in der Multimediawelt. Working paper. University of Karlsruhe. Karlsruhe 1997. Baumol, William: The transaction demand for cash: An inventory approach. Quarterly Journal o f Economics. (Nov) 1952. pp. 545-556 Berentsen, Alexander Digital Money, Liquidity and Monetary Policy. First Monday 2(7). Berkeley 1997. http://www.firstmondav.dk/issues/issue2 7/berentsen/ Berger, Allen N.; Hancock, Diana and Marquardt Jeffrey C.: A framework for Analyzing Efficiency, Risks, Costs and Innovations in the Payments System. Journal of Money, Credit and Banking. 28(4/2) 1996. pp. 696-732 Bemkopf, Mark: Electronic cash and monetary policy. First Monday 1(1). Arlington 1996. http://www.firstmondav.dk/issues/issuel/ecash Bank for International Settlements: Implications for central banks of the development of electronic money. Basle 1996. Blanchard, Olivier J. and Fischer, Stanley: Lectures on macroeconomics. Massachusetts 19961 0 . Brandel, Roland E. and Adams, Lee S.: The modernization of regulation E. ABA Bank Compliance. 18(1). pp. 12-21 Campbell, Colin; Campbell, Rosemary and Dolan, Edwin: Money, banking and monetary policy. New York 1988. United States Congressional Budget Office: Emerging Electronic Methods for Making Retail Payments. Washington 1996. Crede, Andreas: Electronic commerce and the banking industry. The requirement and opportunities for new payment systems using the internet. Sussex 1996. http://www.usc.edu/dept/annenberg/voll/issue3/crede.htm 167 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. [Cuadras 1997] [DePrince 1997] [Dorn 1997] [Eisenbeis 1997] [Ely 1996] [EMI 1994] [EMI 1996] [EMI 1997] [EMI 1998] (England 1996] [Fed 1995] [Fed 1996] Cuadras, Morato: Can ice cream be money? Perishable medium of exchange. Journal of Economics. Zeitschrift fiir Nationalokonomie 66(2). Vienna 1997. pp.103-125. DePrince, Albert E. and Ford, William F.: The Privatization of Currency and Seigniorage. Business Economics. 32(1) 1997. pp. 25-32 Dom, James A.: The future of money in the information age. Washington 1997. Eisenbeis, Robert A.: Bank deposits and credit as sources o f systemic risk. Economic Review. Federal Reserve Bank of Atlanta. 82(3). pp.4-19 Ely, Bert: Electronic money and monetary policy: Separating fact from fiction. In: Dom, James A.: The future of money in the information age. Washington 1997. pp. lOlff. http://www.cato.org/monevconf/14mc-2.html European Monetary Institute: Report to the Council of the European Monetary Institute on Prepaid Cards by the Working Group on EU Payment Systems. Frankfurt am Main 1994. European Monetary Institute: Payment Systems in the European Union. 2n d Blue Book. Frankfurt am Main 1996. European Monetary Institute: Developments in EU payment systems in 1996. Frankfurt am Main 1997. European Monetary Institute: Payment systems in the European Union. Addendum incorporating 1996 figures. Frankfurt am Main 1998. England, Catherine: Cyberbanking and currency competition. In: Dom, James A.: The future of money in the information age. Washington 1997. pp. 137-152. http://www.cato.org/monevconf/14mc-3.html Blinder, Alan S.: Statement to the U.S. House Subcommittee on domestic and International Monetary Policy of the Committee on Banking and Financial Services, October 11, 1995. Federal Reserve Bulletin. 81(12) 1995. pp. 1089-1093 Federal Reserve Bank: A view from the Fed. TMA Journal Atlanta. 16(5). p.20f 168 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. [Fed 1997] [Frankel 1996] [Friedmann 1969] [Gilbert 1996] [Goppl 1996] [Goldfinger 1996] [Good 1997] [Grigg 1996] [Hagel 1997] [Haq 1996] [Hayek 1976] [Humphrey 1996] (TuKDG 1997] Federal Reserve Bank of Chicago: Money Matters. Chicago 1997. Frankel, Jeffrey A.: How well do foreign exchange markets function: Might a Tobin tax help? NBER working paper 5422. Cambridge 1996. Friedmann, Milton: The optimum quantity of money. In: Friedmann, Milton: The Optimum Quantity o f Money and other essays. Chicago 1969. p.39 Gilbert, R. Alton: Financial regulation in the information age. In: Dom, James A.: The future of money in the information age. Washington 1997. pp.71-80. http://www.cato.org/monevconf/14mc-8 .html Goppl, Hermann: Finanzierung. Karlsruhe 1996. Goldfinger, Charles: Electronic Money in the United States: Current status, prospects and major issues. Brussels 1996. http://www.ispo.cec.be/infosoc/eleccom/elecmonev.html Good, Barbara: Electronic Money. Working paper 9716. Federal Reserve Bank of Cleveland. Cleveland 1997. http://www.clev.frb.org/research/workpapr/1997/ Grigg, Ian: The Effect of Internet Value Transfer Systems on Monetary Policy. London 1996. ftp://wueconb.wustl.edu/econ- wp/mac/papers/9607/9607004.html Hagel, John: Net gain. Boston 1997. Haq, Mahbub ul; Kaul, Inge and Grunberg Isabelle (eds.): The Tobin tax: coping with financial volatility. New York 1996. Hayek, Friedrich A.: Denationalization o f Money. Salzburg 1976. Humphrey, David B. and Pulley, Lawrence B.: Cash, Paper, and Electronic Payments: A Cross-Country Analysis. Journal of Money, Credit and Banking. 28(4) 1996. pp. 914-939 Informations- und Kommunikationsdienste Gesetz (IuKDG) (Information and Communication Services Act) passed by the German Bundestag on 13 June 1997 (BGB1.1 S .1870) 169 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. [Jordan 1996] [Keller 1998] [Kocherlakota 1996] [Lengwiler 1997] [Levey 1997] [Loshin 1997] [Lynch 1996] [Madamas 1996] [Marimon 1997] [Matonis 1995] [Me Andrews 19971 ] [Me Andrews 19972 ] Jordan, Jerry L. and Edward, J. Stevens: Money in the 21st century. In: Dom, James A.: The future of money in the information age. Washington 1997. pp. 115-126. http://www.cato.org/monevconf/14mc-6.html Keller, Roland: Immer mehr Firmen setzen auf das virtuelle Geschaft. Handelsblatt. 5/16/1998. Munich 1998. Kocherlakota, Narayana and Wallance, Neil: Optimal allocations with incomplete record keeping and commitment. Working paper 578. Federal Reserve Bank of Minneapolis. Minneapolis 1996. Lengwiler, Yvan: A model of money counterfeits. Journal of Economics. Zeitschriftfiir Nationalokonomie. 65(2). Vienna 1997. p. 123 Levey, Marc M.; O’Donnell, Thomas A. and Powers, J. Pat: Cyberspace transactions present interesting international, state and local tax issues. Tax Executive. 49(6). p.476ff Loshin, Pete and Murphy, Paul: Electronic commerce: online ordering and digital money. Rockland 19972. Lynch, Daniel and Lundquist, Leslie: Digital money: the new era of Internet commerce. New York 1996. Madams, Nick: Options in electronic payment systems. Telemarketing & Call Center Solutions. 15(3). pp.76-80 Marimon, Ramon; Nicolini, Juan P. and Teles, Pedro: Electronic Money: The end of inflation? Discussion Paper 122. Federal Reserve Bank of Minneapolis. Minneapolis 1997. Matonis, John: Digital cash and monetary freedom. Moss Beach 1995. http://www.publitronix.com/digital cash/home.html McAndrews, James J.: Making payments on the Internet. Business Review. Federal Reserve Bank of Philadelphia. Jan/Feb 1997. pp. 3-14 McAndrews, James J.: Network issues and payment systems. Business Review. Federal Reserve Bank of Philadelphia. Nov/Dec 1997. p.l5ff 170 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. [Meyer 1986] [Miller 1966] [Moore 1991] [Nixon 1996] [Obstfeld 1983] [Panurach 1996] [Peirce 1995] [Philips 1996] [Radigan 1997] [Ridley 1994] [Roberds 1993] [Roberds 1997] Meyer, Paul: Money, financial institutions and the economy. Homewood 1986. Miller, Merton H. and Orr, Daniel: A Model of the Demand for Money by Firms. Quarterly Journal of Economics (80) 1966. pp.413-435 Moore, Linda: Money in the third millennium. In: Solomon, Elinor H.: Electronic money flows. Boston 1991. p.61ff Nixon, Brian: E-cash. America's Community Banker. 5(6). p.34f Obstfeld, Maurice and Rogoff, Kenneth: Speculative Hyperinflations in Maximizing Models: Can We Rule Them Out? Journal of Political Economy 91(4). Chicago 1983. pp. 675-687 Panurach, Patiwat: Money in electronic commerce. Digital cash, electronic fund transfer and Ecash. Association for Computing Machinery. Communications of the ACM. 39(6). pp.45ff Peirce, Michael and OMahony, Donal: Scaleable, Secure Cash Payment for WWW Resources with the PayMe Protocol Set. Dublin 1995. http://www.w3 .org/Conferences/WWW4/Papers/228/ Philips, Jim: Bytes of Cash: Banking, computing and personal finance. First Monday 1(5). Arlington 1996. http://www.firstmondav.dk/issues/issue5/ecash Radigan, Joseph: Locking up the money monopoly. USBanker. 107(1). pp. 26-34 Ridley, M.: Electronic money: So much for a cashless society. The Economist. 333(7891). pp. 21f Roberds, William: The Rise of Electronic Payments Networks and the Future Role of the Fed with Regard to Payment Finality. Federal Reserve Bank of Atlanta Economic Review. 78(2) 1993. pp. 1-22 Roberds, William: What’s really new about the new forms of retail payment? Federal Reserve Bank of Atlanta Economic Review. 82(1) 1997. pp. 32-45 171 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. [Samuelson 1980] [Santomero 1996] [Sargent 1993] [Schmeck 1996] [Schreft 1997] [SEC 1997] [Solomon 19911 ] [Solomon 19912 ] [Solomon 1997] [Stuber 1996] [Tatsuo 1996] [Tobin 1956] [Tracey 1997] Samuelson, Paul A.: Economics. Hamburg 1980. Santomero, Anthony M. and Seater, John J.: Alternative Monies and the Demand for Media o f Exchange. Journal of Money, Credit and Banking. 28(4/2) 1996. pp. 942-960 Sargent, Thomas: Rational Expectations and Inflation. New York 19932. Schmeck, Harmut: Informatik B, WS96/97. Karlsruhe 1996. ftp://ftp.aifb.uni-karlsruhe.de/pub/V orlesuneen/Winter1996- 97/Info-B/vorlesungsfolien/ Schreft, Stacey L.: Looking Forward: The Role for Government in Regulating Electronic Cash. Federal Reserve Bank of Kansas City Economic Review. 82(4) 1997. pp. 59-84 Securities and Exchange Commission: Annual Report of First Virtual Holdings Inc. 1996. Commission File Number 000- 21751 (Form 10-K). Washington 1997. Solomon, Elinor H.: Today’s Money: Image and Reality. In: Solomon, Elinor H.: Electronic money flows. Boston 1991. Solomon, Thomas and Solomon, Elinor H.: Money stability and control: The perverse effect of feedback loops. In: Electronic money flows. Boston 1991. Solomon, Elinor H.: Virtual Money. Oxford 1997. Stuber, Gerald: The Electronic Purse: An Overview of Recent Developments and Policy Issues. Technical report 74. Bank of Canada. Ottawa 1996. Tatsuo, Tanaka: Possible Economic Consequences of Digital Cash. Tokyo 1996. http://ftp.uni- mannheim.de/ftp/info/inet 96/papers/bl/bl l.htm Tobin, James: The Interest Elasticity of the Transactions Demand for Cash. Review o f Economics and Statistics 38. pp.241-247. Tracey, Brian: Banking on the Net: Bankers Rethink Opposition To Regulating Electronic Cash. American Banker. 162(82). p.6:l 172 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. [Treasury 19961 ] [Treasury 19962] [Wayner 1996] [Wenninger 1995] [Wenninger 1996] Department of the Treasury: Toward electronic money and banking: The rule of the government. An introduction to electronic money issues. Washington 1996. Department of the Treasury: Selected tax policy implications of global electronic commerce. Washington 1996. Wayner, Peter: Digital cash: commerce on the net. London 1996. Wenninger, John and Laster, David: The electronic purse. Current Issues in Economics and Finance 1(1). Federal Reserve Bank of New York. New York 1995. Wenninger, John and Orlow, Daniel: Consumer payments over open computer networks. Research paper 9603. Federal Reserve Bank of New York. New York 1996. 173 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. IMAGE EVALUATION TEST TARGET (Q A -3 ) A / < • v , A 1.0 £1 l.l 2.8 I 4.0 2 - 5 2.2 2.0 1 . 8 1.25 1 .4 1 .6 150mm IIW 1GE . In c 1653 East Main Street Rochester, NY 14609 USA Phone: 716/482-0300 Fax: 716/288-5989 0 1993. Applied Image. Inc., All Rights Reserved Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Debt reduction by way of inflation: The case of Lebanon

PDF

An enquiry into the effectiveness of export assistance organizations serving southern California SMEs

PDF

Belize pegged exchange rate: Heading towards devaluation or maintaining credibility

PDF

Decentralizing of public finance: Centralizing forces in developing countries

PDF

A study concerning the relevance of using quality measures of education in economic growth models of sub-Saharan African nations

PDF

A model of student performance in principles of macroeconomics

PDF

Does intense probation monitoring of truants work? An empirical econometric analysis of its effect on school and individual attendance, grades, and behavior

PDF

Economic valuation of impacts to beneficial uses of water quality in California: Proposed methodology

PDF

An analysis of SME export assistance needs

PDF

Business group affiliation in India

PDF

In search of the sun: Yukio Mishima's body aesthetic

PDF

Explanations, the availability heuristic, and biases of subjective likelihood in economics

PDF

Intragroup evaluations, attitude source, and in-group member derogation

PDF

Constructions of Americanism: Three case studies

PDF

A prolegomenon to a Los Angeles arts and culture magazine

PDF

Effects of threat and self-focus on consensual bias in majorities

PDF

Everything Went Silent and Suddenly the Grey Mattered

PDF

Essays in supply chain management

PDF

Effects of personal resource sufficiency on perceived difficulty and desirability of earthquake preparedness

PDF

Education in Athens Co., Ohio: A clash of cultures

Asset Metadata

Creator Strasser, Georg Heinrich (author)

Core Title Digital money

School Graduate School

Degree Master of Arts

Degree Program Economics

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag business administration, banking,Economics, Finance,Economics, General,OAI-PMH Harvest

Language English

Contributor Digitized by ProQuest (provenance)

Advisor [illegible] (committee chair), [illegible] (committee member)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c16-26163

Unique identifier UC11337094

Identifier 1393188.pdf (filename),usctheses-c16-26163 (legacy record id)

Legacy Identifier 1393188.pdf

Dmrecord 26163

Document Type Thesis

Rights Strasser, Georg Heinrich

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA