Abstract: This paper examines the role of cash usage in the development of payment media. For that purpose, the basic trends of change in payment media are scrutinized using European data. Although cash still clearly dominates in small retail payments, there has been a definite shift towards electronic

payments in larger purchases and payment of wages and state benefits. At the same time, the distribution of cash has been moving from bank branches into ATMs. This paper focuses on the role of ATMs and ATM networks in dispensing cash. It is shown that ATMs have a very different effect on banks and consumers, which makes it difficult to predict future developments in ATM networks. During the past decade the number of ATMs has increased rapidly, but there are signs of saturation. The number of ATMs is not independent on the overall usage of cash in society. As for the use of cash, it also depends on choices that are made with respect to the ATMs, but demand for cash appears to be continuing to grow in absolute terms. To be able make predictions on the future course of development, we have to assess how such properties of cash as anonymity compare with speed and security. There has been no systematic assessment of this to date.

Key words: Cash, ATM, payment media, seigniorage.

**********

During the last two decades, payment media have changed a great deal. Along with economic growth and the deepening of financial markets, various technical innovations have created new possibilities for settling payments (BOUNIE et al, 2006, for example). Somewhat surprisingly, cash continues to be one of main payment instruments, and in some countries the main instrument (1). Cash has dominated small retail payments, especially in terms of the number of transactions, and the role of cash as the only legal means of payment has supported its status. So far, cash is also the only means of payment that does not require any investment in settlement devices. Technical innovations have also made the use of cash much easier. The number of ATMs has continued to increase and, while the number of bank branches has started to decrease, the total number of service points has increased. Transaction costs for consumers have consequently fallen. Even with the growth in the number of ATMs, the number of ATM networks has decreased. In many countries there is only one network, so that all ATMs can be accessed with the different banks' debit cards. Transportation of notes and coin has been rationalized; handling and authenticity sorting of banknotes has benefited from several technical innovations; and institutional arrangements are greatly lowering the costs of using cash compared to two decades ago. On the other hand, some technical and institutional improvements have made other payment media more competitive vis-a-vis cash.

When considering the use of cash and the choice of payment media, we often hear quite conflicting arguments from different interest groups. The central banks, other banks, shop-keepers and consumers have an interest in the development of payment media, including the tariff structure. Some other players such as security firms, cash-in-transit companies, producers of technical equipment and insurance companies have also joined in the debate. Finally, we have to acknowledge the interest of governments and (in Europe) the European Union.

In the case of cash, there is typically a difference of opinion between banks and the central bank, and consumers. Banks quite naturally see the use of cash as a cost item only. The use of cash lowers the level of deposits, and thus the interest income. Moreover, banks must pay most of the handling and transportation costs. If all people used debits cards only, banks would save a lot in running costs and investment expenses. The central bank and consumers, on the other hand, seem to benefit from the current system. The central bank gets its seigniorage revenue, which is essential to its existence. As for consumers, they probably find the current system optimal because they use cash when it is cumbersome or even impossible to use debit or credit cards or other cashless payment media (in small transactions, person-to-person transactions, and transactions where the nature of the cash payment is important) and other payment media for larger transactions, repeated payments, advance payments etc. We should not forget that cash has some unique properties such as, for example, immediate settlement, independence of payment systems and anonymity, which certainly has increased the popularity of large-denomination banknotes.

Here, we do not participate in this debate (2), but instead illustrate how payment media have developed and how the use of cash has changed, as a part of this process. In particular, we scrutinize the manner in which technical innovations, like ATMs, have contributed to this end.

In what follows, we briefly review some basic trends in the development of payment media and the use of cash. Then, we discuss the dual role of ATMs. The last section presents some concluding remarks.

* Development of payment media

Developments in payment media can be followed easily via, for example, the ECB Blue Book on payment systems and individual countries' bankers' association publications. Here we only refer to a couple of graphs that illustrate the most recent trends in this development.

[FIGURE 1 OMITTED]

Figure 1 shows EU15-wide trends in the use of cashless payment instruments. The message conveyed by the figures is quite clear. The use of cards, credit transfers and direct debits has increased steadily, while the use of cheques has declined to almost one half from the 1992 level. The most rapid growth has taken place in card payments, which increased almost fivefold in 1992-2005. Comparing the use of cash and other payment media is a bit difficult. Some idea of the market shares of different payment media can be obtained from various consumer surveys (such as VIREN, 1994), but we also mention here that in 2005, ATM withdrawals in euros exceeded the card payments in the euro area. In the whole of the EU15, the opposite was true, but both items were still of the same magnitude.

One might expect the use of cash to have diminished considerably in response to the strong growth of alternative electronic means of payment. However, if one scrutinizes figure 2, which illustrates the development of cash balances in Finland and the euro area, an immediate reaction is that there has not been such a marked decrease, not even in the ATM banknotes used for ordinary retail purchases. That becomes even more obvious if we scrutinize the development of different denomination banknotes since January 2002 (figure 3). The demand for euro (area) banknotes has grown considerably, both before the euro conversion in 2002 and since then.

[FIGURE 2 OMITTED]

[FIGURE 3 OMITTED]

There are at least three major reasons for this. Firstly, the demand for euro cash has increased due to increased tourism and cross-border exchanges. Withdrawals of cash in other than the home country may be expensive, so that cash is usually withdrawn at home if one is travelling within the euro area. Secondly, there is evidence that euro cash has replaced the U.S. dollar in hoarding currency, especially since exchange rate expectations have favoured the euro. Thirdly, the use of euro cash replaced the U.S. dollar and other currencies outside of the euro area. euro currency is a much more international currency than the previous national legacy currencies, and the euro is already the most used cash currency in the world. Based on bulk transfers of cash, the estimated amount of euro cash outside the euro area is believed to be 10-20 percent of total euro value currently in circulation.

The demand for the largest denomination, the EUR 500 banknotes, has increased particularly strongly, which suggests that demand for euro cash is related to purposes other than cash usage in ordinary retail payments (3). That also shows up in the fact that large denominations have lower return frequency rate to the central bank than, for example, ATM banknotes or change banknotes. It is often argued that the demand for EUR 500 banknotes is explained by its use in the "grey markets". Examples of such might be transfers of interest income across borders, real estate purchases for tax avoidance, and wage payments to illegal workers. High-denomination banknotes may be used in, but are not the reason for illegal activities. In addition there are also numerous reasons for using cash in entirely legal situations (such as the purchase of cars or other equipment) (4).

Here we are not able to settle this question of why the use of cash has grown; we can only refer to some recent analyses of the reasons for increased currency demand (cf. FISCHER et al, 2004). The magnitude of the recent increase becomes clear if one scrutinizes the case of Finland, where the use of cash has typically been lower than in most other countries. Thus, the most recent currency ratio, at the end of 2006, was the same as in the mid-1960s--long before all card payments (figure 4)!

[FIGURE 4 OMITTED]

One may ask whether--in addition to the possible reasons mentioned above--recent growth can be explained by development of ATMs (or more generally, changes in the number of ATMs and bank branches). That is something that this paper tries to evaluate below. Of course, we cannot completely separate ATM developments from the use and distribution of cash as a whole. For banks and also for customers, ATMs have simply been a convenient and inexpensive way to dispense cash. However, there are signs that ATM usage has already peaked. In the banking community, the prevailing attitude is that less cash should be used in transactions and, in general, the use of (debit) cards should be increased.

[FIGURE 5 OMITTED]

[FIGURE 6 OMITTED]

[FIGURE 7 OMITTED]

The reason why ATMs are so important is simply that the number of ATMs has increased enormously (figure 5) in almost all countries except for Finland (figure 6). Thus, nowadays, there are typically more ATMs than bank branches (figure 7). That, in turn, has changed the total number of service points, which has lowered the transaction cost of cash. The idea that the transaction costs have declined can be defended with the observation that ATMs enable substantially longer opening hours, although over-the-counter services in a bank office may involve even lower transaction costs (besides the fact that bank branches typically include many service points). The importance of service hours becomes clear if we look at the history of ATMs (see the boxed text above).

* The role of ATMs

Empirical analyses have typically found the effect of ATMs on cash demand to be ambiguous (STIX, 2003 and VIREN, 1992, for example). The reason is obvious: ATM density has two conflicting effects on optimal cash holdings. On the one hand, an increase in ATM density lowers transaction costs, which in a typical Baumol-type transaction model (BAUMOL, 1952), leads to smaller cash withdrawals and smaller cash holdings. At the same time, however, cash becomes more competitive vis-a-vis other payment media--say card payments--which increases the share of transactions in cash and that, in turn, tends to lead to higher average cash holdings. Thus, when modelling the relationship between cash holdings and number of ATMs, the result depends on the way transaction technology and costs are modelled: are there ATM fees, and what kinds of fees are charged for debit and credit cards/card transactions? Still it appears that the probable outcome will be that, if the number of ATMs increases cash holdings will increase as well (SNELLMAN & VIREN, 2006) (5).

Given the demand function for cash, one can derive the optimal number of ATMs from a bank's optimization problem. But the outcome depends crucially on the market structure. Thus in a monopoly model, the result is almost trivial (irrespective of the cash demand function): ATMs represent only costs for a bank and this simply leads to a corner solution with respect to the number of ATMs. The optimal number of ATMs is then simply zero (6). Obviously, the only way out this puzzling result is to assume some amount of competition, which forces banks to have a nonzero number of ATMs. In competitive markets banks have to provide some cash services to attract customers who want to use cash. In a simple case in which banks' market share depends on the number of ATMs with a convex market share function, we end up with a relationship between profits and the number of ATMs for a single bank as is illustrated in figure 8. There will consequently be some optimal level of ATMs which balances the costs of cash (for banks) and the benefits from larger market shares (7).

[FIGURE 8 OMITTED]

Thus, if we turn things upside down, we might expect to find a positive relationship between the number of banks and the number of ATMs. In fact, the number of banks may not be the relevant indicator, but rather the number of bank networks. For instance, in the case of Finland, all banks jointly formed an Automatia company, which handles all cash services (ATMs plus money transportation and handling). Automatia is obviously in a monopoly position. Not surprisingly, Automatia has decreased the number of ATMs steadily since 2001 (figure 5). Although there is no indication that Automatia would radically reduce the ATMs in the future, it is clear that banks have a strong incentive to do so. Even if banks would otherwise compete with others, they could separate the cash and transaction business from other business activities (MATUTES & PADILLA, 1994). In any case, as can be seen from figure 9, the number of ATM networks has decreased over time, so that now in most European countries there is only one ATM network.

[FIGURE 9 OMITTED]

Still there are important differences between European countries, for instance in terms of the menu of notes that can be obtained from ATMs (table 1). If the ATMs provide only large denomination notes, so that the minimum withdrawal is large, that represents a kind of implicit transaction cost for the consumers, which may reduce the use of cash. Thus far, explicit transaction costs/fees have not been widely used (McANDREWS, 2001). It is not clear why. Perhaps setting fees for the use of ATMs would also require fees for over-the-counter withdrawals in bank branches, and that is considered overly costly for image reasons (8). It should be noted here that ATM fees would not necessarily lower cash demand because they could lead to the deployment of more ATMs (HANNAN & BORZEKOWSKI, 2006).

From a social point of view, the question is whether we should actively try to reduce the use of cash in transactions. Banks have strongly argued that cash is a "too costly" or "inefficient" means of payment (See McKINSEY, 2006, for example). Although various numbers have been put forward, there is still a lack of balanced evidence on the pros and cons of cash. Competition issues also deserve some attention. If transactions can be effected by just one means of payment the pricing of that means of payment could pose a problem, especially in a monopolist banking sector (CARLTON & FRANKEL, 1995).

* Conclusions

The payment media have changed a great deal during last two decades. This change has been mainly dictated by technical innovations, not so much by tariffs and prices. In the future, things may change if the banking sector moves towards a system where the profitability of all individual activities and products is more systematically monitored. Unfortunately, we know too little about the impact of tariffs and prices on the demand for different media to be able to predict what these kinds of changes might ensure.

Another issue requiring further research is the determinants of cash demand. As we have seen, there are many puzzling features of cash demand that are probably not merely related to changes in transaction technology, but also to macroeconomic and public economy considerations. If we could build a model that properly combines transactions technology and other explanations for changes in cash demand, the "cash exposure puzzles" could perhaps be solved.

These considerations suggest that it is difficult to predict future developments in payment media. It is clear that new technical innovations, such as mobile phones, will affect demand patterns. They may not, however, affect cash demand so much, but rather the use of cards and other "electronic" payment media. Along with SEPA (Single European Payment Area), some changes in payment patterns will probably take place, and these changes will affect the use of cash. They will also speed up the process of abandoning cheques, which are probably the most inefficient means of payment. Thus, the efficiency of the payment system as a whole will certainly increase.

References

BAKER D. (1995): "Shared ATM networks--the antitrust dimension", Federal Reserve Bank of St. Louis Review, November/December, pp. 5-17.

BALTO D. (1995): "Payment systems and antitrust: can the opportunities for network competition be recognized?", Federal Reserve Bank of St. Louis Review, November/December, pp. 19-40.

BAUMOL W. (1952): "The transactions demand for cash: An inventory theoretic approach", The Quarterly Journal of Economics, Vol. 66, No. 4, pp. 545-556.

BOUNIE D. & FRANCOIS A. (2006): "Cash, Check or Bank Card: The effect of transaction characteristics on the sue of payment instruments", Telecom Paris, Working papers in Economics and Social Sciences ESS-06-05.

CARLTON D. & FRANKEL A. (1995): "Antitrust and payment technologies", Federal Reserve Bank of St. Louis Review, November/December, pp. 41-54.

DREHMANN M. & GOODHART C. (2000): "Is cash becoming technologically outmoded? Or does it remain necessary to facilitate "bad behaviour"? An empirical investigation into the determinants of cash holdings", Discussion Paper 358, LSE Financial Markets Group, Discussion Paper Series.

DREHMANN M., GOODHART C. & KRUEGER M. (2002): "The challenges facing currency usage: will the traditional transaction medium be able to resist competition from the new technologies?", Economic Policy, Vol. 17, No. 34, pp. 195--227.

FISCHER B., KOHLER P. & SEITZ F. (2004): "The Demand for Euro Area Currencies: Past Present and Future", ECB Working Paper 330.

GOWRISANKARAN G. & KRAINER J. (2006): "The Welfare Consequences of ATM Surcharges: Evidence from Structural Entry Model", NBER Working Paper 12443.

HANNAN T. & BORZEKOWSKI R. (2006): "Incompatibility and Investment in ATM Networks", Federal Reserve Board, Staff Working Paper 2006-36.

HIRVONEN J. & VIREN M. (1996): "The use of cash in Finnish business firms", Bank of Finland Bulletin, No. 3, 3-9.

MATUTES C. & PADILLA P. (1994): "Shared ATM Networks and Banking Competition", European Economic Review 38, pp. 1113-11138.

McANDREWS J. (2001): "Automated Teller Machine Network pricing--A Review of the Literature", Review of Network Economics 2, pp. 146-158.

McANDREWS J. & ROB R. (1996): "Shared ownership and pricing in a network switch", International Journal of Industrial Organization, 14, pp. 727--745.

McKINSEY (2006): Payment Services in the Netherlands: an Analysis of Revenues and Costs for banks, McKinsey & Company Inc, The Netherlands.

MOOSLECHNER P., STIX H. & WAGNER K. (2006): "How are payments made in Austria? Monetary Policy & the Economy", Quarterly Review of Economic Policy Q2/06, Oesterreichische Nationalbank.

SNELLMAN H. & VIREN M. (2006): "ATM Networks and Cash Usage", Bank of Finland Discussion Paper 21/2006.

STIX H. (2003): "How do debit cards affect cash demand?", Working Paper 82, Oesterreichische Nationalbank.

VIREN M.:

--(1992): "Financial innovations and currency demand, some new evidence", Empirical Economics, pp. 451-461.

--(1994): "Demand for different payment media in Finland", Bank of Finland Bulletin, No. 2, pp. 12-16.

WHITESELL W.:

--(1989): "The Demand for currency versus debitable accounts: Note", Journal of Money, Credit and Banking, Vol. 21, No. 2, pp. 246-251.

--(1992): "Deposit banks and the market for payment media", Journal of Money, Credit and Banking, Vol. 24, No. 4, pp. 483-498.

ATMs slowly reach the end of their life cycle

ATMs were invented 40 years ago by John Barron in the UK to provide customers round-the-clock cash services. One of the triggering facts behind the need was that Europe was moving towards a 5-day working week and banks were no longer open on Saturdays. Originally ATMs were operated for 10 pound cheques using of a 4-digit PIN code. ATMs initially operated as off-line machines and cheques were cleared later within a national clearing system. At first the possibility of getting cash anytime was regarded as a competitive advantage for a banking group. That is why banks in the UK and USA quickly followed the suit. The success and rapid increase in the number of ATM was first related to the fact that Barron's ATM was not patented. Later on, patent applications were in fact filed in some other companies. The reason for this was that patented, it would have revealed too much information about the security features used, therefore security risks were not increased. In time more advanced versions of ATMs entered the market with on-line connections via plastic cards and with the facility of dispensing cash in irregular amounts. It has been estimated that there are currently about 1.4 million ATMs operating in the world, so it will take a while before ATMs are replaced by cashless point of sales (EFTPOS).

In the EU area the number of ATMs is still increasing as ATMs are replacing banking staff and moving cash dispensing from counters to machines. Originally, ATMs were invented to save the costs of cash distribution for banks, but now ATMs are also considered as a means of generating revenues from transactions. ATMs have proven to be more popular than expected for over 40 years. Originally they were seen simply as an intermediate phase in moving to electronic means of payment (from John Barron's interview "The Invention of the ATM", in Currency News April 2005, Vol. 3, No. 4")

Kari TAKALA

Bank of Finland

Matti VIREN

University of Turku (Finland) and Bank of Finland

(*) Useful comments from Alistair Milne, David Mayes and David Bounie are gratefully acknowledged. Financial support from the Yrjo Jahnsson Foundation for Viren is also gratefully acknowledged.

(1) For instance, based on survey accomplished in 2005 in Austria 86% of households' direct payment transactions and 70% of total payment value were cash payments. See MOOSLECHNER, STIX & WAGNER (2006).

(2) See VAN HOVE (2007), in this Dossier.

(3) Currently the share of 200 and 500 EUR denomination banknotes in the euro area is 39%. Adding the 50 and 100 EUR notes, the share goes up to 89 percent. Figure 6 gives some idea of the development of ATM money in Finland, which is probably used mainly in transactions.

(4) See DREHMANN & GOODHART (2000) and DREHMANN, GOODHART & KREUGER (2002) for analysis of the role of cash in modern society. See HIRVONEN & VIREN (1996) and VIREN (1994) for analysis of cash demand by households and firms.

(5) It is not at all clear how to model the demand for cash. Perhaps the most straightforward way is to use a spatial model where transaction costs are related to density of bank branches/ATMs. Alternatively, the key variable would be the size of payments which determines the domain of transaction of a means of payment (WHITESELL, 1989, 1992). In both cases, assumptions on tariffs and prices appear to be quite decisive.

(6) The result would change only if a bank wants to lower its costs by transforming branch offices into ATMs. But this is simply an answer to the problem of minimization of cash handling costs, and we would have no explanation of why banks altogether have any service points for cash withdrawals.

(7) See BAKER (1995) and BALTO (1995) for a deeper analysis of the competition issues.

(8) Except the work of GOWRISANKARAN & KRAINER (2006), very little is known about the effects of fees and prices.

Table 1--ATM banknotes in euro area countries from the start
of changeover

                 500 [euro]        200 [euro]        100 [euro]

Austria                                                   x
Belgium
Finland
France
Germany                                                   x
Greece
Ireland
Italy
Luxembourg                                                x
Netherlands
Portugal
Spain

                  50 [euro]         20 [euro]

Austria        [x.sup.(2004)]
Belgium               x                 x
Finland               x                 x
France                x                 x
Germany               x                 x
Greece                x                 x
Ireland               x                 x
Italy                 x                 x
Luxembourg            x                 x
Netherlands           x                 x
Portugal              x                 x
Spain                 x                 x

                  10 [euro]         5 [euro]

Austria               x
Belgium                                 x
Finland
France                x
Germany               x                 x
Greece
Ireland               x
Italy                 x
Luxembourg            x                 x
Netherlands           x                 x
Portugal              x                 x
Spain                 x                 x