Transaction Pricing and the Adoption of Electronic Payments: A Cross

You don't have the latest version of Adobe Flash Player.

Please update your flash player.

Please login or register to make a comment!

Transaction Pricing and the Adoption of Electronic Payments: A Cross-Country Comparison 7 Wilko Bolt, a David Humphrey, b and Roland Uittenbogaard a a De Nederlandsche Bank b Florida State University After safety, the e ciency of a nation 9s payment system is a primary concern of central banks. Since electronic payments are typically cheaper than paper-based or cash payments, pric- ing these transactions should speed up the shift to electron- ics. But by how much?

Norway explicitly priced point-of-sale and bill-payment transactions and rapidly shifted to electronic payments, while the Netherlands experienced a similar shift without pricing. Controlling for terminal availability and dif- ferences between countries, direct pricing accelerated the shift to electronics by about 20 percent. The quid pro quo was the elimination of bank-Qoat revenues.

JEL Codes: D12, G21. 1. Introduction The average bank cost of an electronic payment is one-third to one- half that of its paper-based equivalent or cash (cf.

Humphrey et al. 2006). A merchant 9s average cost of accepting a bill payment elec- tronically over a giro network or at the point of sale (POS) is also lower (credit cards excepted).

Since the resource cost of a country 9s payment system may account for 1 32 percent of ... more. less.

its GDP, it is clear 7 The authors would like to thank Hans Brits, Nicole Jonker, Rein Kieviet, Raymond Kleijmeer, Andrew Stone, Carlo Winder, and two anonymous refer- ees for valuable comments and suggestions. The views expressed in this article are those of the authors alone and do not necessarily represent those of De Nederlandsche Bank, the European System of Central Banks, or the Federal Reserve Bank of Philadelphia. Author e-mails: Bolt: w.bolt@dnb.nl; Humphrey: dhumphrey@cob.fsu.edu; Uittenbogaard: r.a.uittenbogaard@dnb.nl. 89 90 International Journal of Central Banking March 2008 that shifting from paper to electronic payments can confer social benePts. Importantly, the discounted value of these cost benePts will be larger the more rapidly this shift occurs. Additional e ects also exist and are of concern to central banks since the replacement of cash by electronic payments can alter the monetary aggregates (Duca and VanHoose 2004), reduce government seigniorage revenues (Humphrey, Kaloudis, and Øwre 2004), and make tax evasion and illegal transactions more di cult to hide. There is overwhelming evidence that consumers respond to price incentives but very little evidence on how strong this response may be in the payments area. Although consumers are used to respond- ing to price incentives, they tend not to welcome the opportunity to trade o perceived payment preferences with relative prices when their payment use has commonly been viewed as being cfree. d While businesses often pay directly for the payment services they use via explicit transaction fees or compensating balances, consumers have traditionally paid implicitly through lost Qoat or lower (or no) inter- est on transaction balances. However, consumer surveys indicate that certain implicit costs (e.g., availability, convenience, and secu- rity) can a ect payment choice (Borzekowski and Kiser 2006; Klee 2006), and models have been developed to discriminate between pecuniary and behavioral reasons for credit card versus debit card use (Zinman 2005). These and other motivations for choosing dif- ferent payment instruments were outlined in a recent survey, but their relative importance has not been determined (Scholnick et al. 2007). As per-transaction pricing of consumer payments is rare in Europe and the United States, these analyses cannot address what the impact of explicit pricing would be. Such pricing is rare since, reportedly, banks fear a loss of deposit market share if they are the Prst (and only) bank to implement it, while antitrust authorities would be suspicious of industry e orts to coordinate the implementa- tion of per-transaction prices to minimize changes in relative market shares. One country 4Norway 4has overcome these di culties by coor- dinating only the timing of when direct pricing of consumer pay- ments would start 4not the level of prices to be charged, which could in fact be zero. The quid pro quo was a phasing out of banks 9 practice of recouping payment costs through payment Qoat 4debiting con- sumer accounts prior to a value date for bill payments or delaying Vol. 4 No. 1 Transaction Pricing 91 funds availability for credits to accounts 4which made it appear that payment use was cfree d because the monetary cost to users was indi- rect and implicit. The goal was to make payment costs more explicit so consumers could match better the benePts and costs of di erent payment instruments, a response expected to lower the social cost of their payment system (Enge and Øwre 2006). Our purpose is to determine the e ect of di erential transaction- based pricing of payment instruments on the adoption rate of electronic payments. This is done by comparing the shift to elec- tronic payments in two countries 4one that has transaction pricing (Norway) and one that does not (the Netherlands). Transaction- based prices are key since they directly a ect consumers 9 decisions about payment use, whereas implicit prices and Pxed fees can have limited behavioral e ects since these costs do not vary with usage and, when imposed, are typically stable over time. Nonprice e ects 4 such as availability, convenience, and security 4also inQuence pay- ment use, and our attempt to control for these e ects utilizes a two-country model. Data on payment instrument use for many developed countries is available annually in various Bank for International Settlements and European Central Bank documents, as well as from payment sta- tistics by national central banks. As these time series rarely exceed Pfteen years, a parsimonious model speciPcation is necessary. A com- parable time series of actual payment instrument prices on a broad range of payment instruments is available only for Norway. We con- trast the rapid adoption of electronic payments in Norway over 1990 3 2004 with the experience of the Netherlands, which also rapidly adopted electronic payments but did not impose per-transaction prices on consumers. By applying a system estimation to our model, we improve the degrees of freedom and increase the e ciency of our estimators. If the incremental e ect of direct pricing is large, holding con- stant other intracountry and cross-country inQuences a ecting the adoption of electronic payments, then the potential social benePt can also be large. This suggests that antitrust concerns raised by possible bank coordination of the implementation of prices (but not the level) could be o set by subsequent social benePts. We Pnd that while pricing has speeded up the shift to electronic payments, the shift was only about 20 percent faster in Norway with pricing than in 92 International Journal of Central Banking March 2008 the Netherlands without pricing. Our analysis extends earlier work on payment pricing (Humphrey, Kim, and Vale 2001) by separating terminal-availability inQuences from price e ects for point-of-sale transactions and by doubling the time period covered to capture more price variation. Newly available data also allows us to analyze the paper/electronic trade-o for giro-based bill payments as well as issues associated with nonstationarity in aggregate trended pay- ment data. Finally, we control for unspeciPed and hard-to-measure nonprice inQuences to obtain potentially more accurate price elasticities. In what follows, section 2 illustrates how the composition of payments has evolved in Norway and the Netherlands along with the levels of relative paper and electronic prices in Norway. Explicit transaction prices for consumers are zero in the Netherlands. Our focus is on the substitution of debit cards for cash (or cash and checks) at the point of sale along with the substitution of electronic giro payments for paper-initiated giro transactions. In section 3, a parsimonious ccountry-di erence d model is spec- iPed to separate the e ect of pricing on point-of-sale debit card use and ATM (automated teller machine) cash withdrawals from dif- ferences in terminal availability and real personal consumption in our two countries. A similar model relies primarily on prices for the substitution of electronic versus paper-initiated giro payments (Internet connections far exceed giro use and are not a constraint, and, in any case, an electronic or paper giro can be initiated by phone or in person). No good data exists on payment use prior to the start of pricing (1986) in Norway, so nonprice attributes of dif- ferent payment instruments can a ect measured price elasticities if standard analysis is applied using only one country (as was the case in Humphrey, Kim, and Vale 2001). The Netherlands, which did not price, is used to chold constant d nonprice attributes of the di er- ent payment instruments to obtain potentially more accurate price e ects. Put di erently, we seek to csubtract d the shift to electronic payments that presumedly would have occurred without pricing 4 due to nonprice attributes, terminal availability, and per-person con- sumption levels 4from the shift that is observed with pricing plus these three inQuences on use. Our set of four equations is estimated in section 4 in a seemingly unrelated regression framework to improve e ciency. The e ects of Vol. 4 No. 1 Transaction Pricing 93 prices on payment composition, including the implied price elastici- ties, are presented here. Di erent models are estimated to judge the robustness of the price e ect under alternative speciPcations, such as di erent lagged relationships, Prst di erences, and error correc- tion. A summary of our results is contained in section 5, along with an estimate of the bank cost savings associated with the shift to electronic payments. 2. Payment Composition, Pricing, and Other In<uences on Payment Instrument Use 2.1 Payment Composition Both Norway and the Netherlands experienced a relatively rapid change in their payment composition for point-of-sale and bill- payment transactions over 1990 32004. Point-of-sale instruments are now almost solely debit cards and cash, but in the early 1990s, checks were also important. 1 As seen in table 1, the number of debit card transactions per person per year in Norway rose from 5 to 146 over our Pfteen-year period, growing 25 percent per year. 2 The Netherlands started from a smaller base of one transaction per per- son per year but rose to seventy-seven, a 33 percent annual growth. 3 1 Unlike in the United States, credit card use is low in Europe (and miniscule in the two countries we examine). 2 Oil company terminals and cards were introduced in the 1980s as a substi- tute for cash at gas stations. Although these terminals also accepted bank debit cards, oil company cards could not be used elsewhere and were not priced. The Norwegian payment statistics do not include oil company transactions as debit card purchases (Norges Bank 2000, 33) and neither do we. Oil company termi- nals are included in our series of debit card terminals, however, since they accept debit cards for payment. 3 Checks written per person in Norway went from twelve per person annually in 1990 to less than one in 2004. In the Netherlands, they went from seventeen to zero. Credit card transactions per person in both countries were less than one in 1990 and only three per person (the Netherlands) to Kve (Norway) in 2004 (or about 3 percent of card use in each country). The dominance of debit cards over credit cards is probably due to the fact that banks 4not the credit card companies 4through a joint venture were the Krst to introduce EFTPOS (elec- tronic funds transfer at point of sale) directly from deposit accounts and have POS terminals connected to the bank network installed in shops. The banks 9 pur- pose was to replace checks and cash with cheaper electronic cards at the point of sale. 94 International Journal of Central Banking March 2008 Table 1. Payment Instrument Use Per Person in Norway and the Netherlands (1990 32004) 1990 1993 1996 1999 2002 2004 Growth Rate Debit Card Transactions Norway 5 16 36 71 113 146 25% Netherlands 1 4 24 44 66 77 33% ATM Cash Withdrawals Norway 14 17 22 24 23 22 3% Netherlands 8 20 26 28 30 28 9% Electronic Giro (Credit Transfers + Direct Debits) Norway 15 18 29 46 65 78 12% Netherlands 44 54 70 93 116 124 7% Paper Giro (Credit Transfers) Norway 53 44 52 38 24 18 2 7% Netherlands 34 32 33 27 21 18 2 5% Source: DNB statistics, www.dnb.nl; Norges Bank Annual Report on Payment Systems, www.norgesbank.nl. Part of this di erence is due to the fact that Norway started 1990 with more debit card terminals in place than the Netherlands and so was at a higher point of usage on their logistic growth curve. In 2004, the average amount of a debit card transaction was about e 55 in Norway and e 44 in the Netherlands. No country has time-series data on the number of cash transac- tions, although a few (markedly) di erent estimates exist for some countries at di erent points in time. These estimates di er primarily because of the di culty of estimating very small-value cash trans- actions in which coins are often used and for which stored value cards 4which so far have very limited acceptance 4are the only real substitute. 4 We use the number of cash withdrawals at ATMs as 4 Brits and Winder (2005) provide an estimate of cash use in the Netherlands in 2002. Cash accounted for 85 percent of POS transactions and 56 percent of Vol. 4 No. 1 Transaction Pricing 95 our indicator of cash use in transactions. 5 This quantity measure seems appropriate since the average real value of an ATM with- drawal rose by less than 1 percent annually in both countries. 6 While each cash withdrawal ( e 138 on average in Norway and e 107 in the Netherlands in 2004) funds multiple actual cash transactions, the act of withdrawing cash is priced in Norway, while its use at the point of sale is not. Thus, we compare debit card and cash use at the point where both are actually priced and consumer choice is exercised. 7 Analysis of the U.S. market suggests that the price depositors pay when withdrawing cash from a foreign ATM (an ATM owned by a di erent bank) a ects ATM deployment decisions (to earn rev- enue) and tends to tie depositors to banks with larger ATM net- works (Hannan 2005; Knittel and Stango 2004). These e ects are not important in the Netherlands, as consumers are not charged for ATM usage from any bank, and they play a minor role in Norway where there is only a foreign ATM fee set by the depositor 9s bank but no surcharge by the ATM owner. Thus, the contemporaneous determination of ATM price and terminal deployment is weak, and, in any event, we lag terminal availability in our model speciPcation to mitigate this possible endogeneity problem. sales value, while debit cards accounted for 13 percent of transactions and 40 percent of sales. The average value of a cash transaction was around e 10 but over e 47 for debit cards in 2002. 5 Only recent time-series data exist on the number or value of cash withdrawals over the counter at Knancial institutions or through ccash-back d opportunities at supermarkets. ATM cash withdrawals are the only consistent data that exist for our time period. 6 In 2004 prices, the average real ATM withdrawal in Norway rose gradually from e 127 in 1991 to e 138 in 2004, while in the Netherlands it rose from e 94 in 1991 to e 107 in 2004. These Kgures imply annual real-growth rates of roughly 0.6 percent and 0.9 percent for, respectively, Norway and the Netherlands. There is thus little reason to specify both the number and the value of ATM cash withdrawals. 7 In reality, consumer payment choice is more complex. First, at an ATM there is the choice of whether to withdraw or not; then second, at the point of sale, there is the choice of whether to use cfree d cash or a priced debit card. We leave this ctwo-stage decision d issue aside in our analysis. Our view is that the use of cash at the point of sale will be inLuenced by the cost of consumers 9 replenishing their inventory of cash via an ATM (or other sources). 96 International Journal of Central Banking March 2008 In both Norway and the Netherlands, debit card use expanded at a rapid rate, while growth of ATM cash withdrawals was much smaller. As shown below (and in table 2), the average price of an ATM withdrawal rose relative to a debit card transaction in Norway, but these two prices were both zero in the Netherlands. If relative price was the only inQuence on relative use, we would expect a slower growth for ATM withdrawals in Norway (where the ATM price was, after 1996, higher than debit cards) than in the Netherlands (where there is no di erence in relative prices). We see indications of this for ATMs in table 1 (as transaction growth in Norway is slower than in the Netherlands), but we do not see it for debit card use (where relative prices would favor more rapid growth in Norway over the Netherlands). In order to reQect the substantially lower cost associated with electronic bill payments, employee disbursements, and interbusiness transactions over giro networks, the price of an electronic giro pay- ment in Norway was less than a paper-initiated giro transaction (either delivered in the mail or passed over the counter at a bank or postal o ce). Since giro prices were zero in the Netherlands, one would expect to see a more rapid growth of electronic giro transac- tions and slower growth (or greater reduction) of paper giro transac- tions in Norway than in the Netherlands. Both of these expectations are realized in table 1. Per-person use of electronic giro payments in Norway rose from 15 transactions to 78 over 1990 32004 (growing 12 percent a year) while only rising from 44 transactions to 124 over the same period in the Netherlands (growing 7 percent annually). 8 At the same time, paper giro use fell in both countries but from a higher level and at a greater rate in Norway. Indeed, by 2004 individuals in both countries initiated only 18 paper giro transactions per year. 2.2 Payment Prices in Norway The average 4and sometimes weighted-average 4per-transaction prices being charged for di erent payment instruments in Norway are illustrated in table 2. Since there are no per-transaction fees in 8 By 2004, direct debits accounted for 10 percent of electronic giro payments in Norway but 56 percent in the Netherlands. This is the main reason why electronic giro payments per person in the Netherlands are higher than in Norway. Vol. 4 No. 1 Transaction Pricing 97 Table 2. Average Per-Transaction Prices for Di erent Payment Instruments in Norway Growth Prices in Euros 1990 1993 1996 1999 2002 2004 Rate Debit Card Price Norway . 18 . 23 . 25 . 26 . 28 . 26 2% ATM Cash Withdrawal Price Norway . 05 . 18 . 24 . 29 . 39 . 40 14% Relative Price: Debit Card/ATM Cash Withdrawal 3 . 60 1 . 28 1 . 04 . 90 . 72 . 65 2 11% Electronic Giro Price Norway . 10 . 18 . 22 . 23 . 31 . 27 7% Paper Giro Price Norway . 35 . 62 1 . 18 1 . 86 2 . 65 2 . 76 15% Relative Price: Electronic Giro/Paper Giro . 29 . 29 . 20 . 12 . 11 . 09 2 8% Source: DNB statistics, www.dnb.nl; Norges Bank Annual Report on Payment Systems, www.norgesbank.nl. the Netherlands, the relative prices that Norwegian consumers face also reQect the di erence in prices faced between Norway and the Netherlands. This is the price e ect we wish to separate from other inQuences on payment choice in these two countries. 9 9 Consumers in the Netherlands do face a Kxed annual fee for maintaining an account (about e 6) that allows the holder to use a debit card and the bank 9s ATM network to withdraw cash, while Internet banking typically has a one-time startup fee (around e 15). Because these fees are Kxed, consumers csee d a zero price per additional transaction and respond accordingly. If the Kxed fees, which did not vary much over time, were added as additional variables in our (dou- ble log) model, basically only the intercept and not the price elasticities would be a ected. Additionally, while the average value of a debit card transaction was between e 40 and e 50, about 3 percent of all Dutch debit card transactions 98 International Journal of Central Banking March 2008 The weighted-average per-transaction price of a cash withdrawal in Norway was in 1996, and earlier, less than that for debit cards. 10 This was because a cash withdrawal at one 9s own bank was free during business hours and prices applied only to withdrawals after business hours or at another bank 9s ATM. While debit cards started out in 1990 with a price that was more than three times higher than the weighted average of di erent ATM prices (table 2, row 3), it ended up being only 65 percent of the cash withdrawal price in 2004. Thus, only after 1996 did the absolute price of a debit card favor its use over cash when EFTPOS terminals were available. 11 But even before 1996, there was an indirect inducement to use debit cards in Norway when it became possible in late 1992 to obtain ccash back d from a debit card transaction at the point of sale. 12 This avoided the extra cost and inconvenience of having to use an ATM to withdraw cash, since small amounts of cash could be obtained at no additional cost when making purchases at the local market. There was a stronger relative price inducement to use an elec- tronic rather than a paper-initiated giro transaction for consumer bill payments. In 1990, the price of an electronic giro transaction was only 29 percent as high as a paper giro payment, but by 2004 were subject to a merchant surcharge of about 15 eurocents (De Nederland- sche Bank 2004) when the transaction value was less than e 10 3 e 12. Some mer- chants wished to discourage use of debit cards for low-value transactions since accepting cash is cheaper due to bank fees paid by merchants for debit card transactions. 10 This observation only holds on a per-transaction basis. On average, one ATM withdrawal could fund roughly two to three debit card transactions. How- ever, since this di erence in ctransaction domain d between both instruments is relatively stable over time, it should only a ect the intercept in our model in logs. 11 The relative debit card/ATM price changes shown reLect banks 9 initial e orts to induce depositors to shift cash withdrawals from branch o ces to cheaper ATMs and then later from cash use to even cheaper debit card transactions. 12 Although cash-back transactions and cash at the counter at one 9s own bank are also sources for obtaining cash for free in Norway 4and implicitly lower the e ective price for obtaining cash compared to our use of the weighted average of free and priced ATM access 4these data are available only for recent years and therefore could not be directly included in the analysis. The alternative of including a cash-back dummy, since it was collinear with the debit card price variable already in the model, yielded anomalous results, suggesting that these two e ects cannot be reliably separated. Thus, the debit card price elasticity reported below is best considered as a combination of price and cash-back e ects in Norway. Vol. 4 No. 1 Transaction Pricing 99 this had fallen to only 9 percent of the paper price. In the begin- ning, electronic giro payments were initiated via telephone, but this was later overtaken by the spread of Internet banking. This applies to credit transfers where the consumer retains control in initiating a payment, as opposed to a direct debit where the receiver of the credit initiates the debit to the consumer 9s account under a pre- arranged contractual agreement. Billers often give a slight discount to customers who pay by direct debit, thus creating a slight price advantage over a credit transfer. However, regardless of which party initiates the payment, both are counted as a single electronic giro transaction. It is important to note that the prices charged in Norway do not cover the full bank cost of making a payment (cf. Flatraaker and Robinson 1995; Gresvik and Øwre 2003). In 1988, transaction prices covered only around 25 percent of the banks 9 payment cost, but this coverage had risen to around 70 percent in 2001. 13 As well, in both countries banks initially made some e ort to inform customers of the advantages of using lower-cost electronic payments whether or not the transactions were directly priced (but we cannot explicitly account for this in our model). 2.3 Terminal Availability and Levels of Consumption While relative prices provide an inducement to use electronic pay- ments at the point of sale, this can be accomplished only if a mer- chant has an EFTPOS terminal that can be used. This observation points to the two-sided nature of the payment market, which inQu- ences the adoption rate of new payment instruments. In particular, the market for electronic payment services is considered a two-sided market in the sense that both consumers and merchants are needed simultaneously to demand and cconsume d card payments. Suppliers of payment card services (or so-called cplatforms d) can e ectively cross-subsidize between merchants and consumers through di eren- tial pricing to stimulate this demand. In two-sided markets, typi- cally only one side is charged on a per-transaction basis, while the 13 The relationship between fees and underlying costs is di erent in Sweden, with surplus bank revenues from card transactions cross-subsidizing the expense of providing cash, distorting resource allocation (Sveriges Riksbank 2004, with more detail in Guibourg and Segendor 2007). 100 International Journal of Central Banking March 2008 other side obtains the service (almost) for free in order to generate greater demand. 14 Indeed, merchants value a wide di usion of pay- ment cards among consumers, while consumers benePt from high terminal density at retail locations that accept their cards. In our analysis, payment card and ATM terminal density are included to take this two-sided e ect into account in explaining relative payment card usage. Table 3 shows the number of EFTPOS terminals in place in Norway and the Netherlands over 1990 32004 per one million of Table 3. Terminal Availability, Real Consumption, and Demographic In<uences on Payment Instrument Use Growth 1990 1993 1996 1999 2002 2004 Rate Debit Card EFTPOS Terminals (per mil population) Norway 2,487 6,324 8,932 13,214 17,723 21,091 15% Netherlands 148 1,600 6,170 9,176 10,941 11,967 34% ATM Terminals (per mil population) Norway 419 396 426 451 484 473 0.8% Netherlands 180 291 395 421 465 468 6.6% Real Per Capita Personal Consumption (in 1,000) Norway 11.9 12.1 13.9 15.1 17.8 16.7 2.3% Netherlands 9.2 9.3 9.9 10.9 11.4 11.3 1.4% Share of Young Adults in Population Norway 8.0 7.8 7.2 6.4 6.0 6.0 2 1.9% Netherlands 8.5 8.2 7.0 6.1 6.0 6.0 2 2.3% Source: DNB statistics, www.dnb.nl; National Accounts; Dutch CBS; Norges Bank Annual Report on Payment Systems, www.norgesbank.nl; IFS. 14 In Norway, the consumer side is directly charged for its use of payment instru- ments, while in the Netherlands, the retailer side of the market pays per transac- tion. Bolt and Tieman (2004) provide an explanation for these widely observed completely skewed pricing strategies in two-sided markets. See Rochet and Tirole (2003) for a rigorous analysis of two-sided markets and competition. Vol. 4 No. 1 Transaction Pricing 101 population (which controls for di erences in population size). 15 As shown in the Prst two rows, Norway had almost twice as many debit card terminals as the Netherlands in 2004, and this di erence was far more extreme in earlier periods. While the growth of EFTPOS terminals has been more than twice as rapid in the Netherlands, it still has a long way to go to provide the same density of termi- nal access as Norway. By this measure alone, it really would not be possible 4regardless of any price incentive 4for consumers in the Netherlands to use debit cards with the same intensity per person as they do in Norway. As noted earlier, there is no price incentive to use debit cards in the Netherlands, so there are two reasons 4no price incentive and fewer EFTPOS terminals per person 4to expect that the Netherlands would use debit cards less intensively than Norway. Even so, as shown below, it is di cult to separate the e ect of prices from terminal availability on debit card and ATM use. The same cseparation problem d exists for cash withdrawals at ATMs. Norway prices ATM withdrawals, while the Netherlands does not, and for the entire period Norway also provided a greater den- sity of ATMs to withdraw cash from (table 3, row 3). Separating the price e ect from the terminal e ect for ATM cash withdrawals may be somewhat easier here since by 2004 both countries had almost the same ATM density but withdrawals were priced only in Norway and, compared to the Netherlands after 1993, per-person use in Norway was correspondingly less (table 1, row 3). 16 Inferences on the relative importance of pricing may be more accurate if two other possible, but small, inQuences on payment choice are considered. One concerns di erences in the level of real 15 In 2004, the population in the Netherlands was 16.3 million; in Norway it was 4.6 million. 16 As Norway is roughly nine times larger than the Netherlands, di erences in population density may compromise the usefulness of our availability measure of ATM and EFTPOS terminals. However, both countries are highly urbanized, which is probably the most important driver for installing terminals. In Nor- way, the Kve largest cities account for about 25 percent of total population but only 1 percent of total geographic area (see Norway statistics, www.ssb.no). Less extreme, in the Netherlands, the ten largest cities make up roughly 20 percent of Dutch population, with 3.5 percent of the geographic area (see CBS statistics, www.cbs.nl). Since this di erence in densities is e ectively a constant over Kfteen years, in our log-di erence equation its impact would a ect only the intercept and not the slope parameter, which is our terminal elasticity. 102 International Journal of Central Banking March 2008 per capita personal consumption between the two countries, since higher levels of real consumption tend to be associated with larger numbers of transactions. 17 A second inQuence concerns the possibil- ity that changes in the number of young adults in both countries may a ect di erences in new payment adoption rates. Consumer surveys indicate that young adults and higher-income individuals adopt new payment arrangements more rapidly than others, even without pric- ing. But direct pricing could well a ect the adoption rates of those with greater habit persistence, those with a lower opportunity cost, or those who do not value much the added convenience or security that electronic payments can o er. The level and variation of both per capita consumption and the share of young adults in the population over time are illustrated in the bottom half of table 3. Real per capita consumption in Norway was 29 percent greater than that in the Netherlands in 1990 but rose to be 48 percent higher in 2004. This di erence should be associated with a rising number of all types of transactions in Norway relative to the Netherlands. There are smaller di erences between these two countries in the shares of young adults 4new entrants into the labor force aged twenty to twenty-four. Indeed, these shares are falling in both countries. 18 3. A Country-Di erence Model of Payment Choice Di erences between Norway and the Netherlands are used to try to explain per capita use of debit cards, ATM cash withdrawals, and electronic and paper giro payments. As outlined above, the main inQuences on payment use and composition are di erences in the number of EFTPOS and ATM terminals per million population, the prices being charged in Norway (positive) and the Netherlands (zero), and di erences in the level of real per capita consumption. 17 All monetary values for Norway (prices as well as real consumption) have been translated from Norwegian kroner into euros using a purchasing power parity exchange rate. Also, real per capita consumption in Norway includes oil revenues only indirectly, as some of this revenue is used to Knance government expendi- tures, which likely reduces taxes from what they otherwise would be, permitting real consumption to be larger. 18 Demographic variables are typically extremely smooth series. In implemen- tation, this created convergence problems in our system estimation and the population share variable was excluded. Vol. 4 No. 1 Transaction Pricing 103 Our time period is short (only Pfteen years), as time-series data on payment instrument use have only recently been deemed important enough to be routinely collected at the country level by government agencies. While some time-series data on some payment types do exist for longer periods in some countries, this information is not comprehensive, nor are payment instrument prices available, since very few types of payment services are directly priced. Norway is the exception that allows us to undertake this analysis. These data constraints impose a parsimonious speciPcation on our explanatory four-equation model: CARD t = ± 1 + ± 2 CARDTERMINAL t 2 1 + ± 3 CARDPRICE t + 1 / 2 ± 22 CARDTERMINAL 2 t 2 1 + ± 33 CARDPRICE 2 t + ± 23 CARDTERMINAL t 2 1 7 CARDPRICE t + ± 4 CONSUMPTION t + µ 1 t (1) ATM t = ² 1 + ² 2 ATMTERMINAL t 2 1 + ² 3 CARDPRICE t + 1 / 2 ² 22 ATMTERMINAL 2 t 2 1 + ² 33 CARDPRICE 2 t + ² 23 ATMTERMINAL t 2 1 7 CARDPRICE t + ² 4 CONSUMPTION t + µ 2 t (2) EGIRO t = ³ 1 + ³ 2 EGIROPRICE t + 1 / 2 ³ 22 EGIROPRICE 2 t + ³ 3 CONSUMPTION t + µ 3 t (3) PGIRO t = ´ 1 + ´ 2 EGIROPRICE t + 1 / 2 ´ 22 EGIROPRICE 2 t + ´ 3 CONSUMPTION t + µ 4 t . (4) In the variable dePnitions below, NOR indicates Norway and NL indicates the Netherlands. Di erences between these countries are expressed in index form: 19 19 In many cases, the log of the absolute di erence in our variables between countries was negative (or changed from positive to negative), so all variables are expressed as the log of the ratio or index of the di erence between countries. 104 International Journal of Central Banking March 2008 CARD = ln (NOR debit card use/NL debit card use), on a per-person basis; CARDTERMINAL = ln (NOR card terminals/NL card termi- nals), per million population; CARDPRICE = ln (NOR card price/NOR ATM price); CONSUMPTION = ln (NOR personal consumption/NL per- sonal consumption), real per capita; ATM = ln (NOR ATM cash withdrawals/NL ATM cash withdrawals), per person; ATMTERMINAL = ln (NOR ATM terminals/NL ATM termi- nals), per million population; EGIRO = ln (NOR electronic giro use/NL electronic giro use), per person; EGIROPRICE = ln (NOR electronic giro price/NOR paper giro price); PGIRO = ln (NOR paper giro use/NL paper giro use), per person. Since debit card and ATM terminals have to be in place before consumers can use them 4and, even when in place, typically have a lag before they are used at a signiPcant level 4these two termi- nal variables are lagged by one year in the model to give a closer correspondence between the exogenous availability of new terminals and their possible e ect on use. Prices, of course, also have to be known before they can a ect payment choice. The lag here is likely much shorter, and prices are speciPed as exogenous and contem- poraneous. 20 Note that in the absence of Dutch prices for payment instruments, we implicitly take the relative price for the Netherlands as being constant, which will then only a ect the intercept in our model. Looking at the data, it appears that the two countries di er in when they introduced electronic payment instruments. In the Netherlands, usage of the electronic giro was on a higher level than in Norway in 1990, whereas Norway had a higher density of ATM and EFTPOS terminals. This cstarting value d problem is taken into account in our logarithmic speciPcation through the intercept, which 20 The e ects of di erent lag arrangements on the results are noted in section 4. Vol. 4 No. 1 Transaction Pricing 105 is not restricted to a value of 1 (which would imply equal starting values for 1990). 4. Estimation Results and the E ect of Price on Payment Instrument Use The system of equations (1) 3(4) was estimated in a seemingly unre- lated regression framework to allow for the possible correlation between errors in locally identifying debit card use with those for ATM cash withdrawals and similarly for electronic and paper giro use. With Pfteen observations per equation, there are thirty-eight degrees of freedom (d.f. = 4 7 15 2 22). As shown in the appen- dix, the explanatory power of the model was high (the respective adjusted R 2 s were .96, .98, .84, and .75 from the system estima- tion). As the variables are not I(0) and thus our levels estimation may be unbalanced, yielding spurious results, we Prst checked for the presence of any residual autocorrelation. Fortunately, system resid- ual portmanteau (Ljung-Box) Q-test statistics (adjusted for small sample) indicate that the autocorrelations of the residuals are not statistically signiPcant. As well, the Durbin-Watson values are fairly reasonable for our four levels equations (respectively, 2.12, 2.80, 1.86, and 2.03). These autocorrelation tests suggest that the variables of our four equations are likely to be cointegrated. To formally test for cointegration, we applied augmented Dickey- Fuller (ADF) tests to the residuals of the levels equations (1) 3(4), using critical values computed by Phillips and Ouliaris (1990). The test statistics show that we reject the null of no cointegration at the 5 percent level for the ATM and the electronic giro equation, and at a 10 percent level for the debit card and paper giro equa- tion (see the last column in table 5, discussed later). 21 The signs of the estimated parameters appear to be reasonable and expected from theory, so the degree of spurious correlation, if any, is likely to be small. Moreover, as shown in the next subsection, the residuals of our levels equations 4measuring the deviations of a clong-run d 21 By deKnition, cointegration requires that the variables be integrated of the same order. Keeping in mind our small sample, applying ADF tests 4using MacKinnon (1996) critical values 4indicated that for fourteen out of seventeen variables, the null of a unit root is not rejected. 106 International Journal of Central Banking March 2008 Table 4. Price and Terminal Elasticities Under Di erent Model Speci;cations Lagged Lagged Terminals Separate Prices Terminals and Prices No Lags Own Substitute Debit Card Price E ect 2 . 22 7 7 2 . 48 7 7 2 . 06 2 .19 2 .03 Terminal E ect . 53 7 7 . 57 7 7 . 94 7 7 .49 7 7 ATM Cash Withdrawal Price E ect . 23 7 7 . 31 7 7 . 29 7 7 2 .86 7 7 .69 7 7 Terminal E ect 2 . 16 7 7 2 . 49 7 7 2 . 35 7 7 2 .35 7 7 Electronic Giro Price E ect 2 . 46 7 7 2 . 53 7 7 2 . 44 7 7 .21 7 .10 7 Paper Giro Price E ect . 27 7 7 . 25 7 7 . 33 7 7 2 .03 .03 Note: Starred ( 7 7 , 7 ) values indicate signi-cance levels of 1 percent and 5 percent, respectively. relationship 4signiPcantly a ect the short-run dynamics of the vari- ables. Hence, we feel that these Pndings support the results of our preferred model using levels data. 22 The derivatives of equations (1) 3(4), Prst with respect to rela- tive prices and then with respect to lagged terminal availability, are shown in column 1 of table 4 (with parameters and other statis- tics shown in the appendix). A 10 percent reduction in the price of debit cards relative to an ATM cash withdrawal is associated with a 2.2 percent rise in the relative use of cards in Norway compared 22 Naturally, given the small sample size, unit-root and cointegration tests have reduced power, and strict interpretation of the estimated elasticities should be made with caution. Vol. 4 No. 1 Transaction Pricing 107 Table 5. Price and Terminal Elasticities: Data in Levels and First Di erences Error Cor. Levels Error Cor. 1st Di . & 1st Di . ADF Data Residuals Levels a Data a Test Debit Card 2 3 . 82 æ b Price E ect 2 . 22 7 7 3 . 25 .47 Terminal E ect . 53 7 7 . 82 .75 7 7 Feedback Parameter 2 0 . 93 7 2 . 07 ATM Cash Withdrawal 2 5 . 55 7 Price E ect . 23 7 7 . 32 7 .06 Terminal E ect 2 . 16 7 7 2 . 39 7 7 .47 7 7 Feedback Parameter 2 0 . 84 7 2 . 83 7 7 Electronic Giro 2 4 . 50 7 Price E ect 2 . 46 7 7 2 . 60 7 7 .37 7 7 Feedback Parameter 2 0 . 71 7 2 . 24 Paper Giro 2 3 . 66 æ Price E ect . 27 7 7 . 49 7 7 .11 Feedback Parameter 2 0 . 85 7 2 . 45 7 Note: Starred ( 7 7 , 7 ) values indicate signi-cance levels of 1 percent and 5 percent, respectively; circled ( æ ) values indicate a 10 percent level. a Debit cards and ATMs formed one system estimation, while electronic and paper giros formed another in the -rst-di erenced and error-correction models. b Test based on one signi-cant explanatory variable. with the Netherlands (which has a zero explicit price for both cards and ATMs). At the same time, a 10 percent increase in debit card terminals in Norway relative to the Netherlands is associated with a 5.3 percent rise in debit card use in Norway relative to the Nether- lands. As seen in the table, lagging both terminals and prices by one period doubles the strength of the price response (from 3.22 to 3.48) 108 International Journal of Central Banking March 2008 but does not alter the terminal elasticity. Assuming no lags, how- ever, increases considerably the apparent responsiveness of debit card use to changes in terminal availability 4making it almost one- to-one in percentage terms 4but the trade-o is that it generates a price elasticity insigniPcantly di erent from zero. Over 1990 32004, the price of ATMs in Norway rose relative to that of debit cards. The price elasticity suggests that a 10 percent rise in the relative price of ATMs is associated with a 2.3 percent decrease in relative use. 23 Numerically, this is very similar to the result for the debit card equation, where a 10 percent reduction in the relative price of debit cards gives a 2.2 percent rise in rel- ative use. 24 The ATM terminal elasticity, however, has an unex- pected sign and is negative at its mean. When evaluated yearly, the terminal elasticity is positive over 1990 394, but the negative relationship for the remaining years dominates, giving a negative mean. Looking more closely at ATM use and terminal availabil- ity by year (not shown) suggests that the source of the negative elasticity is that per-person ATM use in Norway reaches a peak in 1998 and then falls, while ATM availability in Norway reaches a peak Pve years later in 2003. Similarly, ATM use in the Nether- lands peaks in 2001, but terminals continue to expand. The apparent explanation for the negative ATM terminal elasticity is that ATM use has reached saturation (due in part to the price disincentive), while terminals are still being added (allowing banks to substitute ATMs for expensive branches, which fell absolutely in the two coun- tries), giving the result that terminals are expanding while use is falling. The estimated price e ects for electronic and paper giro pay- ments conform to expectations since, when the relative price of elec- tronic giro transactions falls 10 percent, relative use of this instru- ment in Norway rises 4.6 percent compared with the Netherlands. Similarly, a 10 percent increase in the relative price of paper giro 23 Since the price ratio used in the ATM equation is the same as that used in the debit card equation 4ln (Norway debit card price/Norway ATM price) 4the negative debit card price elasticity would become a positive elasticity in the ATM equation. 24 The ATM price e ect is larger when both terminals and prices are lagged in the model. Using bank-level data for Spain, Scholnick et al. (2007) also Knd that debit cards substitute for ATM cash withdrawals. Vol. 4 No. 1 Transaction Pricing 109 payments is associated with a 2.7 percent reduction in relative use between the two countries. 25 Once an individual switched to making an electronic giro payment, almost all of their giro transactions were electronic, and electronic payment volume grew by inducing more and more individuals to switch. In contrast, the substitution of debit cards for ATMs was twofold since it involved individuals shifting a progressively larger share of their point-of-sale transactions from cash to cards over time as more terminals became available and rel- ative prices changed as well as inducing more and more individuals to adopt and use cards. To illustrate the robustness of our results, our preferred model in equations (1) 3(4) was respeciPed so that direct debits, which com- prise 10 percent of electronic giro payments in Norway but 56 percent in the Netherlands, were deleted from the electronic giro use (basi- cally leaving only credit transfers). This had almost no e ect on the price results shown in column 1 of table 4. Equations (1) 3(4) were respeciPed again to include checks with ATMs so that both can substitute with debit cards. Checks were important in the early 1990s, had a high price, and their use e ectively fell to zero by 2004. Nothing of substance was changed except that the debit card price elasticity lost signiPcance. 26 Real per capita personal consumption was markedly higher in Norway and growing faster than in the Netherlands. We expected that this would have a signiPcantly positive e ect on expanding rel- ative electronic payment use in Norway. However, the e ect of real per capita personal consumption on payment use was insigniPcant in all four equations. Just as an exercise, equations (1) 3(4) were simpliPed by deleting the squared terminal, squared price, and terminal-price interaction variables. Then the remaining price ratio in each equation (e.g., debit card price/ATM price and electronic giro price/paper giro price) was reexpressed as the log of separate own and substitute price variables 25 Since the same price ratio is used in both the electronic and paper giro equations 4ln (Norway electronic giro price/Norway paper giro price) 4the neg- ative electronic giro price elasticity would become a positive elasticity in the paper giro equation. 26 The price and terminal elasticities were only slightly changed if, instead of estimating equations (1) 3(4) as a single system, system estimation was applied to (1) and (2) and then separately to (3) and (4). 110 International Journal of Central Banking March 2008 for each equation. The resulting own and substitute price elasticities, along with the reestimated terminal e ect, are shown in the last two columns of table 4. Our preferred model (in column 1) is speciPed in ratio form, due to our limited sample, but it is of interest to see the implied own and cross-price elasticities that result from estimat- ing each price elasticity separately. All but one own price elasticity is negative, and three of the four cross-price elasticities are positive (as would be expected for a substitute payment instrument). However, considering that only one negative own elasticity and two positive cross-elasticities were signiPcant, it seems that the price e ects are not very strong. 4.1 Cointegration and Error Correction In the previous section we formally tested whether the variables of our levels equations were cointegrated. These tests indicated that the residuals were stationary and that we could reject the null hypoth- esis of no cointegration. To assess how deviations from the long-run equilibrium 4as captured by the movements of the residuals 4a ect the short-run dynamics of the variables, we also estimated an error- correction (system) model. Since the models in levels, in Prst di er- ences, or in error-correction form are all nested within an cautore- gressive distributed lag d framework, it allows us to test which model Pts the data best. To illustrate, consider the following extension of equation (1), written in an autoregressive distributed lag regression format by adding lagged endogenous and exogenous variables: 27 CARD t = ± + ³ CARD t 2 1 + ´ 1 CARDPRICE t + ´ 2 CARDPRICE t 2 1 + ² 1 CARDTERMINAL t 2 1 + ² 2 CARDTERMINAL t 2 2 + u t . (5) Without a ecting its ability to explain the data or changing the least-squares estimates of the parameters of interest, (5) may be 27 Note that compared with equation (1), the squared variables, interaction terms, and consumption have been excluded. These additional variables could be included without a ecting our illustration. Vol. 4 No. 1 Transaction Pricing 111 rewritten in error-correction form: CARD t = ± + ² 1 CARDTERMINAL t 2 1 + ´ 1 CARDPRICE t + ( ³ 2 1)( CARD t 2 1 2 ± 22 CARDTERMINAL t 2 2 2 ± 33 CARDPRICE t 2 1 ) + u t , (6) where ± 22 = ² 1 + ² 2 ³ 2 1 and ± 33 = ´ 1 + ´ 2 ³ 2 1 denote the long-run elasticities (equivalent to the elasticities in levels equation (1)). In (6) we have an equilibrium relationship describing the short-run dynamics, CARD t = ± + ² 1 CARDTERMINAL t 2 1 + ´ 1 CARDPRICE t + u t , and an equilibrium error, CARD t 2 1 2 ± 22 CARDTERMINAL t 2 2 2 ± 33 CARDPRICE t 2 1 , which measures the deviation from the long-run relationship between the variables CARD t 2 1 , CARDTERMINAL t 2 2 , and CARDPRICE t 2 1 . Consequently, the feedback parameter ³ 2 1 can be interpreted as the proportion of the resulting disequilibrium that is reQected in the movement of CARD t in one period. If the parameter ³ 1 2 1 is negative and signiPcantly di erent from zero, the model in error-correction format cannot be rejected. 28 In this case, long-run equilibrium deviations have a signiPcant impact on the short-run dynamics, which disqualiPes the model in Prst di erences. On the other hand, insigniPcance of the adjustment parameter would favor a Prst-di erence model and implies a cdisconnect d between the short run and long run. This disconnect would then cast doubt as well on the empirical relevance of the long-run relationship (even when the variables are cointegrated). Because our variables in levels are cointegrated, direct estima- tion of equations (1) 3(4) yields csuper-consistent d estimators of the (cointegrating) long-run elasticities. Under cointegration, the residuals from the levels equations can be used to estimate the 28 As a stability condition, the feedback parameter ³ 2 1 needs to be between 0 and 31. 112 International Journal of Central Banking March 2008 error-correction model. Alternatively, one can estimate equation (6) directly, using Prst di erences and lagged level variables, but this will reduce our already limited degrees of freedom even further. The error-correction estimation results of all four equations are shown in table 5, along with the price and terminal elasticities for our preferred levels model from table 4. The second column shows the adjustment parameters using the stationary residuals of the levels equations (1) 3(4). All parame- ters have the right sign and magnitude, and are signiPcant at the 5 percent level, indicating that the short-run dynamics is indeed signiPcantly inQuenced by deviations from the long-run relation- ship. 29 The Prst-di erence results are shown in column 4 of table 5, but the model is rejected. While the terminal elasticities have the expected sign and are signiPcant, this is at the expense of weak results for the price elasticities. Compared with using residuals in the error-correction estimation, we obtain weaker results when we apply direct estimation of the error-correction equations (see column 3). Here the debit card price elasticity is no longer signiPcant, but the other price elasticities have the expected sign and are signiP- cant (even with a reduction in degrees of freedom), although in two cases the feedback parameter was not signiPcant at the 5 percent level. Given our data limitations, the outcomes of the cointegration tests, and the performance of the error-correction model using sta- tionary residuals, these results weakly suggest that the (cointegrat- ing) price elasticities using levels data in equations (1) 3(4) 4our preferred model 4are robust and can be relied upon as long-run estimates. 4.2 Estimation of Electronic for Paper Substitution in Norway The e ect of pricing on payment instrument use is also esti- mated for Norway alone. This approach should give similar results 29 The residual properties of the error-correction estimation using residuals are fairly reasonable with (adjusted) Ljung-Box stats: Q(1) = 24 . 1 (p = 0 . 09), Q(2) = 43 . 1 (p = 0 . 09), Q(3) = 62 . 8 (p = 0 . 07), and Q(10) = 163 . 5 (p = 0 . 41). Vol. 4 No. 1 Transaction Pricing 113 to our country-di erence model if nonprice characteristics that a ect payment use in a country are weak. The speciPcation is linear and simpler than our country-di erence model (due to degrees-of-freedom considerations), and all the data are for Norway: CARDATM t = ± 1 + ± 2 CARDATMTERMINAL t 2 1 + ± 3 CARDATMPRICE t + ± 4 CONSUMPTION t + µ 1 t (7) ELEPAPER t = ² 1 + ² 2 ELEPAPERPRICE t + ² 3 CONSUMPTION t + 2 t , (8) where CARDATM = ln (debit card use/ATM use), on a per-person basis; CARDATMTERMINAL = ln (card terminals/ATM terminals) , per million population; CARDATMPRICE = ln (debit card price/ATM price); CONSUMPTION = ln (personal consumption), real per capita; ELEPAPER = ln (electronic giro use/paper giro use) , per person; ELEPAPERPRICE = ln (electronic giro price/paper giro price). Equations (7) and (8) were estimated in a systems equation framework (with 23 degrees of freedom, d.f. = 2 7 15 2 7). As shown in table 6, the elasticity of substitution between debit cards and cash was 2 . 20 using levels data and 2 . 31 in Prst di erences. A 10 percent rise in the relative price of an ATM cash withdrawal (which reduces the ratio of debit card to ATM prices) is associated with a small (2.0 percent or 3.1 percent) rise in the ratio of debit card to ATM use. If this parameter was 2 1 . 0, then the expenditure shares of debit cards and ATMs would be unchanged, since a 10 percent 114 International Journal of Central Banking March 2008 Table 6. Price and Terminal Substitution for Norway Levels Data First-Di erenced Data Debit Card/ATM Substitution Price E ect 2 .20 7 2 .31 7 7 Terminal E ect .54 7 7 .06 Electronic/Paper Giro Substitution Price E ect .54 7 7 .13 Note: Starred ( 7 7 , 7 ) values indicate signi-cance levels of 1 percent and 5 percent, respectively. relative rise in the ATM price would be exactly o set by a 10 percent decrease in relative ATM use. Since the parameter is less than 1 (in absolute value), the expenditure share of ATMs rises as the price- induced substitution is less responsive than in, say, a traditional Cobb-Douglas framework where the elasticity of input substitution to a price change is 1.0. The elasticity of terminal availability on debit card and ATM use is .54, which indicates that a 10 percent relative rise in debit card terminals leads to a 5.4 percent relative rise in debit card use. The substitution elasticity between electronic and paper giro transactions initially had the wrong sign (at .54) and was signiPcant using levels data in a system estimation. Use of Prst-di erenced data did not alter this sign, but dropping real personal consumption from (7) and (8), which had a large and signiPcant e ect, did (giving a signiPcant 31.98 value for electronic/paper giro substitution). Over- all, the card/ATM price and terminal elasticity results were simi- lar between our two-country and single-country applications. This occurs because terminal availability appears to be a good instru- ment for the combined e ect of non-price inQuences. In contrast, for giro transactions where such a proxy is not available, the two- country giro price elasticity indicated a much smaller degree of price substitution than the single-country speciPcation, suggesting that nonprice e ects in a single-country model overstate the bill-payment price elasticity. Vol. 4 No. 1 Transaction Pricing 115 4.3 The E ect of Pricing on Payment Use Our overall conclusion is that the availability of card terminals is a good proxy for the net e ect of nonprice inQuences (such as con- venience) on card use. And although nonprice inQuences on card use appear to have a stronger e ect than does per-transaction pric- ing, the shift to electronic payments is speeded up when pricing is present. 30 If both prices and terminals are expanded at similar per- centage rates, then the adoption of electronic payments could have been speeded up by perhaps 40 percent compared with not having per-transaction pricing. 31 As seen in tables 2 and 3, however, debit card terminals changed at a much greater rate than did the price of ATMs or the relative prices of cards to ATMs, indicating that in this instance a potential speedup of 40 percent is too high and was not realized. More precisely, for Norway, the average annual growth in debit card terminal density equaled +15 percent, whereas the growth in card price relative to ATMs was 2 11 percent. Given the estimated elasticities in column 1 of table 4, this would predict a relative rise of debit card use over ATMs of 15% × 0 . 53+11% × 0 . 22=10 . 4% from the terminal and price e ects alone. Without any price induce- ments, this increase in usage would be 15% × 0 . 53 = 8 . 0%, suggesting that the substitution process has been speeded up by approximately 2 . 4 / 10 . 4 = 23%, although the realized contribution of pricing to debit card adoption was 2.4 percent a year. 32 Electronic giro payments do not have a terminal constraint, and the inQuence of consumption growth on payment use is not signiP- cant, so only the e ect of pricing is measured in the single-country case, while nonprice inQuences are incorporated in the two-country estimate. The growth of electronic giro relative to paper giro prices 30 Dutch survey results conKrm the relative importance of terminal availability for payment instrument usage and stress also the nonprice attributes of payment instruments (see Jonker 2007). 31 This estimate is derived from the ratio of the price elasticity in our preferred model ( 322 percent) in column 1 of table 4 to the terminal elasticity (53 percent), which equals .42. 32 The same calculation using price ( 2 . 20) and terminal (.54) elasticities for Norway alone from table 6 gives a 20 percent speedup for debit card use (with a contribution of 2.2 percent annually). 116 International Journal of Central Banking March 2008 was 2 8 percent, while the price elasticity in table 4 was .46, sug- gesting that the realized contribution of pricing to the adoption of electronic giro payments was 8% × 0 . 46 = 3 . 7% annually. Thus, in terms of both the size of the estimated price elasticities and their realized impact on adoption rates, the e ect of pricing on the shift to electronic payments is greater for giro transactions than for debit cards. 5. Summary and Conclusions Electronic payment instruments (credit cards excepted) are consid- erably cheaper than their paper-based alternatives, including cash. Banks and merchants are interested in shifting users to electronic payments to save costs, as are some government policymakers who seek to improve the cost e ciency of their nation 9s payments system. Historically, banks have recouped their payment costs through (i) interest earned on payment Qoat (from delaying availability of funds credited to accounts and debiting accounts prior to bill-payment value dates), (ii) maintaining a spread between market rates and the rate paid on deposits, and (iii) charging Qat monthly fees or imposing balance requirements. In contrast to business users, consumers face very few payment services that are priced on a per-transaction basis and so have little price incentive to choose the lowest-cost instrument either at the point of sale or for bill payments. Banks are well aware that transaction pricing can speed up the shift to electronic payments, but they are reluctant to lose deposit market share by being the Prst (and perhaps only) bank to imple- ment explicit prices di erentiated according to underlying costs. While this problem is mitigated if most (or all) banks implement pricing at about the same time, antitrust authorities are unlikely to view such coordination as being in the public interest unless the social benePts from pricing are signiPcant and the quid pro quo is a compensating reduction in payment Qoat, a higher interest rate paid on deposits, or a reduction in Qat fees or balance requirements. Indeed, Qoat reduction was the trade-o when banks coordinated the timing of when they would implement pricing in Norway (there was no coordination in the prices to be charged, and initially some were zero). Vol. 4 No. 1 Transaction Pricing 117 We use the experience of Norway (which priced its payment ser- vices) and the Netherlands (which did not) over 1990 32004 to try to determine what the incremental e ect of transaction pricing may be on the adoption of debit cards versus withdrawing cash from an ATM and on the adoption of electronic giro transactions (credit transfers and direct debits) versus paper giros. SpeciPcally, we compare per- person payment instrument use in Norway in response to the prices being charged, the availability of terminals, and the level of real con- sumption with the experience of the Netherlands, which also adopted electronic payments but did not price. Our four-equation country- di erence model spanned Pfteen years 4the limit of the available data 4and during this time the share of electronic payments rose by some 60 percentage points, from around the mid-twenties to the mid-eighties, which in most cases easily covered the inQection point where the share of electronic payments switches from rising at an increasing rate to rising at a decreasing rate. Our model is estimated in a systems-equation framework using levels data, and robustness is illustrated by estimating models in a Prst-di erence and error-correction framework. Price and terminal elasticities derived from these models form the basis for our conclu- sions and indicate the incremental e ect of pricing on the adoption rate of electronic payments. The similarity of our card/ATM price and terminal elasticity results between our two-country and single- country applications suggests that terminal availability is itself a good proxy for hard-to-specify/hard-to-measure nonprice attributes of card use at the point of sale. In contrast, for giro bill-payment transactions where such a proxy was not available, the two-country giro price elasticity indicated a much smaller degree of price sub- stitution than the single-country speciPcation. This indicates that unspeciPed nonprice e ects in a single-country framework can over- state the ctrue d giro price elasticity value. The e ects of pricing di er depending on which instruments are being considered. Overall, pricing has a smaller e ect on shifting consumers from ATM cash withdrawals to debit card use than it does in shifting use from paper to electronic giro transactions. The reason for this di erence seems to be that consumers value the non- price benePts associated with debit card use (convenience, security) such that the availability of terminals needed for debit card transac- tions has a stronger e ect on debit card use than prices (as evidenced 118 International Journal of Central Banking March 2008 by the fact that the debit card price elasticity is smaller than the terminal elasticity). Debit cards also substitute for costly checks, and the high price on these instruments in Norway was associated with their virtual elimination, although the same thing happened in

less