Cash, Paper, and Electronic Payments: A Cross-Country Analysis David B. Humphrey, Lawrence B. Pulley, Jukka M. Vesala Journal of Money, Credit and Banking, Volume.28, Issue 4, Part 2: Payment Systems Research and Public Policy Risk, Efficiency, and Innovation (Nov., 1996), 914-939. Stable URL: http:/flinks,jstor.org/sici ?siciA0022-2879% 28 1996] 1%2928%3A4%3C914%3ACPABPA %3E2.0.CO%3B2-N ‘Your use of the ISTOR archive indicates your acceptance af ISTOR’s Terms and Conditions of Use, available at bup:swww,jstor org/abouvtenms.html. ISTOR’s Terms and Conditions af Use provides, in part, thar unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the sereen or printed page of such transmission. Journal of Money, Credit and Barking is published by Ohio State University Press. Please contact the publisher for further permissions regatding the use of this work. Publisher contact information may be obtained at bbupsseww jstor.org/joumalyohio.press:hun). Journal of Money, Credit and Banking (©1996 Ohio State University Press JSTOR and the JSTOR Jogo are vademarks of JSTOR, and are Registered in the U.S, Patent and Trademark Office. For more information on JSTOR contact jstor-info@umich edu. ©2002 JSTOR, up:thvww stor orgy ‘Thu Feb 7 22:23:44 2002DAVID B. HUMPHREY LAWRENCE B. PULLEY JUKKA M. VESALA Cash, Paper, and Electronic Payments: A Cross-Country Analysis A.COUNTRY'S PAYMENT SYSTEM is what makes its real and financial markets work. When commodities are exchanged for cash, check, giro, credit card, or debit card payments—rather than for other commodities —trade ex- ands as transaction costs fall and proxtuction specialization increases. But what determines which payment instrument is used and why are some used ‘more intensively than others? 1s payment instrument use strongly tic to standard demand theory influcnces (such as relative prices and income) or do institutional, habit, and simple payment availabilty differences across countries dominate? What ‘has been the pattern of payment instrument use over time and how similar has this been across counties? Finally, how large are the payment substitution relationships and what does this imply forthe future use of different payment instruments, espe~ ally between paper and electronic payments? ‘To address these questions, we analyze the payment systems of fourteen devel- coped countries using payments data over 1987-1993 recently constructed by the ‘Bank for International Settlements (BIS). In these counties, there were 119 billion fnoneash payments in 1993, an increase of 34 percent over 1987. While 35 peccent of these payments were electronic, this rises to 62 percent when the United States is excluded. Since the cost of an electronic payment ranges between one-third t9 onc~ half that of a check or paper gira payment (Robinson and Flatraaker 1995; Hum- ‘Comments by Robert Avery of the Federal Reserve, Kevin Keamey, Mr. Saasinen of the Bank of Foland, ands leres ae appreciated Author name ders alphabetical aed vefet gus cont ons byallauors. The options expresied ae be suors own nd do aoc ecesanlycepeset tose of he Tank of Filan Davin B. HUMPHREY is professor of finance at Florida State University. LawaBnick B. PULLEY i professor of business administration inthe Graduate School of te College of WE liars and Mary. SURKA M. VESALA 1 ax econamuse in the Financial Markers Department, Ceneral Bark of Finland. Journal of Money, Credit, and Banking, Noi. 28, No. 4 (November 1996, Part 2)DAVID B. HUMPHREY, LAWRENCE B. PULLEY, AND JUKKAM. VESALA = 915 phrey and Berger 1990), the social cost of a country's payment system—which is probably beeween 2 percent co 3 percent of GDP—can be reduced by promoting the shift to electronics. In tis pager we identity the countries that have been most suc cessful in this regard and attempt to detctmine the factors associated with this change.! In what follows, measures of cash use actoss countries are compared in section | and related to the level of all noncash transactions. Cash use and noncash transac tions are negatively related, as may be expected, but—unexpectedly—this implied substitution is due more to differences in use across counties than i isto changes in se over our seven-year period. Section 2 focuses on the relative levels of noncash paper and electronic payments. Although a high percentage of electronic payments is associated with intensive use of cash actass counties, current trends will reverse this relationship in the future: both cash and noncash paper instrument use will be low while electronics will be high. As shown in section 3, in most countries check ard paper giro transactions pet person have reached a peak and ate falling ? Overall, the annual number of electronic transactions per person rose by twenty-four during, 1987-93 while paper-based payments rose by ten. Excluding the United States, the shift to electranies is even more evident: per person electronic payments rose by thirty-three while paper-hased payments rose by three (with check use falling absolutely) A statistical model of the economic and institutional determinants of use of five {types of paper and electronic payment instruments is described in section 4 with results presented in section 5, The analysis frst covers the ten developed countries for which payment instrument price data are available. Time and country dummies are included to capture unspecified time-series and cross-country effects and to ‘gauge the conditional explanatory power of price, income, availabilty institutional, “and lagged use variables. In 2 second estimation, the country dummies are restricted to zero in order to force the coeflicients on the remaining variables to catalogue dif- ferences in payment use across countries. The price variahles (which yield fragile results) are also dropped, permitting us to cover payment use in all fourteen devel- ‘oped countries. Elasticities are computed from the model to indicate the level and significance of substitution among the various payment instauments, as well as be- toveen these instruments and cash, Conclusions are presented in section 6, |. CASH HOLDINGS AND NONCASH TRANSACTIONS, Cash is, by far, the most used payment instrument. The few surveys or estimates that exist place the proportion of cash transactions among all transactions at 90 per- OS ota ay een aed cnn ad mo ues payment we sole amen over rg wt metres seco fo ee Eat pteent of Casein ‘St etcapcly fare vale, and fave bee eles ™ 4. Giro payment, which are common Burope ar cause to preautborzes bil payment the ‘ited Stas Crea for something hs ges around ta orl ane ceerbes the mover of fans rome gr rane 8 par rhe payer ayer, an ly Com epee ack the wo.916 MONEK.CREDIT, AND BANKING cent for the United Kingdom, 86 percent for Germany, 83 percent for the United States and Finland, and 78 percent for the Netherlands (Boeschoten 1992, pp. 73- 74; Humphrey 1984, p. 6; Virén 1993). White data on (otal cash transactions are ‘extremely sparse, {wo inditect indicators of cash use exist and ace shown in Table [ Column | shows the U.S. dallar value of cash held by the public divided by popula- tion in 1993 and is used to rank the fourteen developed countries.” By this measure, ‘each person in Switzerland, Japan, and the Netherlands holds aver $1,300 while ‘each person in the United Kingdom, United States, and Finland holds less than half that amount. The range between the highest user of cash (Switzerland at $2,676 per person) is more than six times that of the lowest (Finland at only $438). ‘Au altemative measure of cash use is the ratio of currency value to GDP, which is shown in column 2. There is good agreement between the three highest and thee lowest users of cash for these two measures and 2 similar—but not exact— correspondence elsewhere. Defiating data on currency per person over 1987-93 by country-specific consumer price indices (to contrat for inflation differences) and re- ‘essing this measure of eeal currency per person on the ratio of currency to GOP. (which aleeady controls for inflation) yields an r of .91. Thus, the two measures are similar indicators of cash use. However, our analysis of noncash payment instra- ‘ments is in terms of annual transactions per person and so we rely on the fist cash measure—cash holdings per person—in the following analysis. “The differences in cash use across Countries seen in Table 1 are related ¢o differ- ences in the publie’s demand to hold cash for transactions, precautionary, and spec lative (hoarding) purposes (Laidter 1985)-+ Cash holdings ner person rise with real per capita income but fall with inflation and higher interest cates (the opportunity cost of holding idle cash balances) * These demands are affected by the extent of illegal actvities— including the avoidance of taxes—across countries. In addition, low crime rates in Japan and most of Europe make it safe for consumers and retailers to.ely more on cash while a high crime rate, which exists forthe United States, has, 4 dual influence: more cash use by the criminal element can be offset by less cash use by consumers and retailers who fear theft.© The availability of automated teller machine (ATM) terminals tends t raise cash use while the availability of credit and debit cards and their terminals tends to reduce. it (Boeschoten 1991, 1992). Debit card payments in particular are close substinuies for cash since both are: used to initiate small-value retail payments. In European bes eer een cone i evince ok seach raryl ei ‘Fett amiga de Mp ty tn indepen mst mcrae ets mange nei sre el Reagent ay + tears al on ng praevia be sal Tee ntl a ade ge 18 ek eet sonia etic 2 Beta tint ra 27 lp 8 pnt oA nat ® Sh eng oa ae SAUER MS RECRLELO Go pte eve ce tyDAVID W. HUMPHREY, LAWRENCE B, PULLEY, ANDJUKKAM. VESALA. 917 TABLET ‘Cast Honpevas ano Annual Noneast TeaNsncions Pur Pony (Fourteen developed counties, Us. §, 1993) Me Saree oem s S coe ino! Sas ae i NGeands C00 649 Switetand Seeun 1310 te ® Newsy tie aa Nerwey ay List ats Began Belg te 335 ue Senay"? 38 is Faint Sera Ms is eam Fame 3 +S Connie Cami 76 Ma Srany ate Kapa oe te Nethrtncs Ute Ste x Fe Foxe Filan Bs eestpeenhy cine a loge of A ae muy hal se oc crs ae a ean a a ‘ie dng (ce Pesan Ttee 185) fhe Seon, me hey he hg raghy wes eid sd Ge x Gry sien amore we Sp exons countries known for theic relative emphasis on card-based payments, such as Fin- land, France, Dentnark, an the United Kingdom, cash holdings per person are cela tively low (Table 1). The expanded availablity of ATM terminals can both raise the overall use of cash in tcansactions (by making, cash withdrawals more corwenient) and atthe same time lower average idle cash balances held by the public (as would bbe predicted from a Batimol-Tobin model with a lower transaction cost for cash ‘withdrawals), The strength of this substitution, as well as the substittion of other payment instriments for cash, can be roughly gauged by comparing cash holdings ‘with the number of all noncash transactions in Table 1 The anmual number of noncash transactions per person is shown in column 3 and is composed of all paper-based payment cransactions (check, paper giro) and all electronic payments (electron giro, eredic card, and debit card point-of-sale (POS) payments)” While there was a six-fold difference in the implied use of cash per person between Switzerland (highest) and Finland (lowest), there isa ten-fold differ- ence in the number of naneash transactions per person berween Italy (the Jowest at ‘swenty-nine per person per year) and the United States (the highest at 292). Impor- tantly, high cash use per person in column I is associated with tow use of noncash, payments in column 3, indicating weak substitution (r ~ ~.79). Icis also of interest 4. Secon gr rnsstions ele det des, dec depois, and tier swore eaing house (ACH payers Maxt ATM ansrtins represen cath wiiawal aac dee payments fo foi deen ats ecac atoncm nme Ine lS eas wera {Eshorcheck depose and acount anaes epee, tapers. S8 pce, cent TATA acta pens ely sen Td Coe he ed Reet Se tem 990)918 | MONEY, CREDIT AND BANKING we? SSS (Cyances Caste Hotnevas aN ANNUAL NONCASHE TRANSACTIONS Pea PERGON OVER 1987-93! ‘Dass Nae, 12g pr Ren ne Op nes Poe og + De 2B a as 6 3 2 3 4 a 3 a 8 i 2 4 4 ae a 0 15 5 4 8 2 “ Fe 3 a ‘arbre Tac ea we ea aad TT to determine the implied elasticity between noncash transactions and use of cash Regressing the log of noncash transactions per person over 1987-93 on the Log of real currency use per person over the same period gives a pooled cross-country time-series elasticity of ~.68, oncaning that a 6.8 percent vise in moncash trans- actions is (sigoificantly) associated with a 10 percent reduction in cash use. (the r= - $64). The relationship between noncash payments and cash use aver time is isolated in ‘Table 2. This table shows the change in the number of annual per person noncash transactions and cash holdings between 1987 and 1993. The countries are ranked by their level of soncash transactions in 1993 (lowest fest)—the same as that for these ‘transactions in Table 1. All fourteen countries experienced a rise in the annual num- ber of noncash transactions per person over this seven-year period.® The average person in the United States, the Netherlands, and France increased their annual nnoncash tcansactions by over forty between 1987 and 1993 while transactions in ‘Sweden, Finland, and ftaly only rose by four or fewer. Bight countries experienced a reduction in. cash use when noncash transactions rose white six experienced an increase in both. Interestingly, the frst group consists of European counties while the second group includes the United States, Canada, Japan, and ltaly. Thus it ap- pears that in Europe the substitution between cash and noncash has been stronger than in the other developed countries.? Although tly (hich had the fewest noncash ransacsions had closet he salle increase ius (four) ae the Ure Sues Cac Fad the tos had the ages ines (Hse) he ise woe ‘uh transactions over 1947-91 wea heterogeaeaus sad aly Weak) cated othe level af noacash Ue tera“ changes me own cama t i pogsesively nee going sown fhe ‘9. The fling sare of cash in MI fam 1983 to 1983 in al f ur counties (ncep aly and pa) implies a he share of payments ue By cash has decease relative wo noneathtansetons Wet rely on depos many In Norway Finland, andthe Uojea Kingdom his eadstie hs cn ele dee ‘rate Cmeratonal Monetary Fund, Pserniional Financia Sut)‘OAV B. HUMPHREY, LAWRENCE B PULLEY, AND JUKKA M. VESALA 913 {In contrast with Table 1 where the cross-country, time-series celationship between rnoncash transactions and cash use is negative and significant (r = ~.79), there is no relationship between these two variables when only the changes in Table 2 are con- sidered (r = ~.01), Since the changes in cash holdings and noncash transactions essentially repcesent only a time-series effect, the negative relationship found in Table | is seen to be primarily due ta differences across countries. 2, NONCASH TRANSACTIONS: PAPER VERSUS ELECTRONICS" For all fourteen countries, there. were seventy-seven billion paper transactions in 1993 and forty-two billion electronic transactions, ! The impression that paper dom- inates electronics is due solely to the pattern of use in the United States. The United States accounts for 64 percent of the 119 billion noncash transactions (but only 36 percent of the population). When the United States is excluded, paper and electronic. transactions umber seventeea billion and twenty-six billion, respectively, showing forthe other thirteen counties that electronics dominates paper. ‘The country by-country payment composition is shown in Figure 1 where each bar shows the annual number of noncash transactions per person in each year over 1987-93. As seen, the number of noncash tcansactions has risen forall countries between 1987 and 1993 (consistent with the uniformly positive changes in column of Table 2). The shaded portion of each bar indicates the number of naperbased transactions per person in each year over the seven-year period. Except for Japan, France, and the United States, paper-based transactions have fallen or are flat over 1987-93. However, the rise in paper transactions in Japan and France has been less than the rise in electronic payments. Electronic payments per person have risen in all fourteen countries over 1987-93 and, with the exception of the United States, this is the primary reason forthe rise in noncash payments, ‘The: comparison between paper and electronic payments ¢an be seen more clearly in Table 3 where the annual number of these two types of transactions are shown for ‘one year (1993), Countries are ranked by the number of noncash payments made per person (lowest firs, as in Table 2). The level of noncash transtctions docs not ap- pear to be related to the (paperfelectronic) composition of these payments. This is evident from the unordered ranking of the petcent of electronic payments in cotunin 3 ‘What is striking, however, isthe vast difference in yearly per person use in paper and electronic payments across countries, Switzerland, Japan, Belgium, and the 10, The cost dierence terween electronic 20d paper-based payrents—noted above—maies the pe elec diction. 0! preter econame raptanee tar an skeraivedstaco, ck a hat ween provaonal and tal eal) payers A provisional Pavmen: sane hat may be revered fo" lexus funds (es check cet rater) while Beal payne tar WOE at cen be Mel unless tye fuss vere suicent, so ee agen eveinls (ke 3 ta ot ACH cre ans). The rovisional-al payeeen distinction is, however, of Bret importance for Wie aster etworks which us om geval payment. Mt Papen praetor omc pape taaton (hth a per na90 MONEY, CRED, AND RANKING eh ca OO a. Anne Nuno Neco, Paper, sn cron Taman eso een developed cutis) Netheslands all initiate fewer than twenty paper instruments annually per person ‘hile the United States initiates over 230, and is clearly an outlier for no other coun- try initiates even 40 percent ofthis figure." There is less dispession regarding elec- tronic payments: ony Jtly initiates less than twenty clectronic payments per person annually while Denmark, Germany, and the Netherlands initiate just over one hundred, ‘Comparing the relative importance of electronic and paper-based payments in all noncash transactions, nine countries initiate 50 percent or more of their noncash payments electronically. Since electronic payments ate the ony source of growth of the nencash payments in eleven countries and the primary source of growth in two ‘of the remaining three (Figure 1)! ics instructive to contrast electronic payments, with the use of cash across countries. The relationship between the percent of elec- tronic payments (Table 3) and currency holdings per person CTable 1) is shown in Figure 2 for 1993 Figure 2 is divided into four quadrants which indicate the spit between electronic and paper-based transactions and how these noncash transactions relate to a coun- try's intensity of cash use, Over all fourteen countries, the average person initiated 12 Checks are he ony important paper insrunent i the Used Sites aos twenty ae wren serio rem Tc setae aca cmben of tie pe y cone St Fetcet),Cosness (4 percent), and government erent, While tales checks ana toe) odes Ute als papers, they el) acount for sro Spetent oY check solune (lupe and Berger 15) and ve bre meget ns my and Bee 13. While France apn, adhe United States ae he dee cous where papas paynents x pznon di ot fal ofr 987-95, he United Sac wa the only sunt sete he owt oper ‘nants per pron (ty one as ager han thf crocs ven sx),DAVID 8. HUMPHREY, LAWRENCE 8 PULLEY, ANDJURKAM. VESALA = 921 TABLES ANNUAL PAPER AND ELECTEONIC TRANSACTIONS Pak PERSON’ Pounzen developed countries, 1995) France rite states 4 » 61 percent of their noncash transactions electronically and held $1,133 in cash (hese figures ate shown as dashed lines which define the quadrants). Figure 2 tells, us that most countries essentially fell into only (wo of the quadrants: they either rely ‘mostly on peper-hased payments and hold relatively litte cash or they tend to se both cash and electronics intensively Countries in the upper right quadrant commonly use cash for small value point- of-sale payments while larger-value bill payments are usually made by elect credit transfers or direct debits. The lower left quadrant is comprised of countries, where checks are an important instrument forall kinds of noncash payments: point of sale, consumer bill payments, and business disbursements, Countries in the up- per left quadeant often rely on plastic card payments in place of cash for small value point-of-sale payments while electronic credit transfers dominate bill payments and disbursements ‘The lack of countries that use hoth paper and cash intensively (lower right quad- rant) indicates that there has been a stronger historical trade-off between paper and ‘cash than there has been, to date, between electronies and cash. However, as shown in Table 4, this is likely t0 change in the future. Table 4 ranks countries by their percent use of electronics (lowest first). Counties with below-average use of elec- tronics, of course, rely on paperbased payments while those with above-average tse of electronics use few paper-based instruments. This table shows the change in paper-based and electronic transactions per person over our seven-year period, along with the corresponding change in cash holdings. Countries which currently have below average use of electronic payinents (op half of Table 4) increased their average por person use of paper instruments by only two payments over seven years 14, Yaragoc (1999) develope sae gure using a crreneyGDP rai in place of real aency ger pean ear renege ery Siar eto ace, aston Te I Ret exes a fave canes92 MONEY, CREDIT, AND BANKING Po, 2, Pecan of Eleni Pats nd Vie of Cureney pr Pere Geen dlp ces, 1960) While the average number of electronic payments per person rose by tweaty-four.'* FFor these countries, the real value af eash holdings rose by an average of $24 per person. However, if Italy is excluded, average per person cash haldings would bave Sallen by $15. Thus itis clear thatthe rise in electronic payments in counties that rely on paper-based payments is associated with a flat or negative growth in per Person use of paper instruments 2s well as of cash. A negative relationship is clearly ‘evident for counties that already have a relatively high-percentage use of electron ies (bottom half of Table 4). Here the average rise of thirty additional electronic transactions per person over 1987-93 is associated with ten fewer paper-based pay- ments and a reduction of $96 in average cash use. . If the (ends showin in Table 4 continue, then i is only a mater of time before the ‘countries in Figure 2 will move up—showing greater use of electronics —and over to the left—showing a reduction in cash use. Although Figure 2 currently suggests a Positive relationship across countries between electronic payments and cash bold ings, consideration of the trend evident within countries suggests that most will move toward the upper left quadrant in the figure. Put differently, electronic pay- iments substitute for both cash and paper-based payments" What is intresting ip this process isthe general uniformity of the expansion of electronic payments pec person across countries, especially that between countries that do and do not now 15. tf the United Sates is exuded, average per pectn eof pape istuners would te fallen by ttc aeetins her an cen yd 16. This takes place maily though he subtion of rei an de card payments for cash in sad sped ats ad te sbattten of ceo prod AT se fsa id paper po payment for ger value ba paymecs and iburemestsDDAVIO B HUMPHREY, LAWRENCE B. PULLEY, AND JURKA M. VESALA. 993 TABLES Cans 1 Pyeranceo ano Buscrnoxtc TRANsAcTONS PR PERSON AND Cast OUTSTANDING ‘onan tf or oF low Average Use of Eocwanis eae) x % se aly 0 : ade) “6 a ‘$ Norway eis -8 a 8 France G32) 4 3 S Une Kirgdom ($05) 4 2 “8 Average: 7 a > Aboverhurage Use of Blacionc Ela ey 2 a1 Sweden 45%) 3 8 Germany C5) is ‘éL “an C5) 4 10 Diem 12) 8 4 Belg ca) S 8 Nether 75) 2 i Sviandand 2%) B 8 ‘Average = Se i a a, era wea VT i a am use electronic payments intensively. This is in contrast to the more heterogeneous experience regarding changes in per person use of cash and paper-based payments. 3. COMPOSITION OF NONCASH TRANSACTIONS Noncash transactions rely om five payment instruments. Paper-based transactions are composed of checks and paper based gito payments. The number of times a per son used cach of these instruments during 1993 is shown in the bar diagram in Fig ure Ja. The lightly shaded (bottom) portion of each bar indicates check use per person per year. As scen, check use dominates use of paper-based giro payments in ‘ight of the countries. This is primarily du. to the fact that many previously paper- based gito payments have already moved to electronics in giro-dominated countries while there has not been a similar shift to electronics in check: dominated countries (the United States, Canada, France, and the United Kingeom Electronic transactions are made up of electronic giro, debit card (POS), an credit card payments.” Per person use of these instruments is shown in Figure 3b. Electronic giro payments chottor portion of each bar) dominate the ather two cate- zories of electronic payments in eleven countries. The three exceptions are Finland 17. Blecronie gio payments clade dt eis (amare bil payments) and ACH wane ios et Seon of pyrll nd sompeate cash maragerae des).9% + MONEY, CREDIT, AND BANKING ham yg 08 Po, Sa, Anal Nantero Paper Tasco ero oun dead stuns) Which eelies on card-based payments and the U.S. and Canada who rely on credit cards (op shaded portion of each bar). In terms of per person use of different payment instruments, che fourteen éevel- oped countries can essentially be divided into «wo groups: the four counties — United States, France, Canada, and the United Kingdom—that vite the most checks per person and the remaining ten chat rely heavily on giro payments (either paper or electronic). Check and all giro payments accounted for 76 percent to 100 Percent ofall noncash payments per person across counties in 1987 and from 68 percent (o 97 percent in 1993. The average reduction in the share of check and giro payments was nine percentage points, a reduction that merely eflets the increasing share of ereit and debit cards in noncash transactions. In almost every country, check and paper giro teansactions per person have reached a peak ant are falling. However, asa total, the annual nuanber of electronic transactions per person rase by twenty-foue during 1987-93 while paper-based pay- ‘ments rose by ten. Excluding the United States, the sift to electronics is even more ‘evident: annual electronic payments per person rose by thiny-three while papcr- based payments rose by three (with check use falling absolutely). ‘These shifts in the use of paper and electronic payment instruments occurs for retail point-of-sale transactions (where credit and debit cards replace cash and checks), consumer bill payments (where electronic giro transfers are replacing ‘checks and paper giro transfers), and disbursements like payroll and other business 16. Chet cad, by roving sho-em cet re more han jst a bute payne wed. In fic United Stas, pent of evo rect ad expends presents tt Non‘DAVID B. HUMPHREY, LAWRENCE B, PULLEY, AND JURKA ML VESALA. 95 ML ‘i Gat tr (8 Fh Arnal Numer of Ecco Tncbas er Pein fous develope cums) payments (which experience the same substitutions as consumer bill payments). De- tailed data on payment instrument use by even these three general transaction cate- gories docs not exist. However, information on the average value per transaction for ‘ur five: payment instruments in Table 5 provides some indication of the purposes for which the different payment instruments are often used. The low average value. per transaction for eredlit and debit cards indicate that these two payment instru- ‘ments are used primarily for retail point-of-sale transactions, not for bill payments ‘or disbursement. We know from practical experience that check and paper and elec- ‘TABLES ‘Aveesor VALUE wen Taansacrion! (Fouten developed counties, US. $, 1995)