`
`Carol L. Clark
`
`Introduction and summary
`During the 1990s, some payment analysts suggested
`that smart cards1 with e-purse applications could be a
`promising new payment option for certain types of trans-
`actions. An e-purse is a stored-value payment device
`that offers the following features to the consumer: It
`holds electronic monetary value that substitutes for
`cash; it does not require online authorization; it records
`the value of each purchase on the card rather than a
`central computer server; and it can be exchanged for
`goods and services from various merchants. The de-
`vice is generally stored on a computer chip, which can
`reside on any one of a number of items most consumers
`already carry, such as a payment card, mobile phone,
`key chain, or even a watch. When the consumer makes
`a purchase, monetary value is deducted from the mi-
`crochip on the card.
`The key difference between a stored-value smart
`card and debit, credit, payroll, and gift cards is that
`value is stored directly on the smart card rather than
`stored in an account on a central computer server, and
`therefore, transactions are processed offline between
`the smart card and the card reader at the point of sale
`(POS). In contrast, debit, credit, payroll, and gift cards
`in the United States are offered on magnetic stripe
`cards, and payment involves an online authorization
`that requires a real-time connection with a central com-
`puter. The purchase is approved or declined through
`the authorization process, which checks whether there
`is sufficient value in the account for debit, payroll,
`and gift card transactions and whether the credit limit
`has not been exceeded for credit card transactions.
`The authorization process may also check whether
`the card is fraudulent or stolen.
`Some payment analysts predicted that smart cards
`could lead to a cashless society, one in which e-purs-
`es would replace cash and coins for low-value pay-
`ments. As we know, this hasn’t happened. Although
`
`a number of e-purse programs have been implement-
`ed around the world, these programs have experienced
`varying degrees of success, and many have failed out-
`right. Smart card adoption in the United States has been
`slower than in the rest of the world. Many analysts argue
`that this is partly because the U.S. already has an ad-
`vanced telecommunications infrastructure that can
`verify magnetic stripe card transactions quickly and
`cheaply online. This results in relatively low fraud
`levels and relatively high levels of satisfaction among
`businesses and consumers with the current systems.
`If this is true, then smart card applications may offer
`more value in other parts of the world with less highly
`developed telecommunications infrastructures and high-
`er incidences of fraud in existing payments networks.
`In this article, I review six e-purse smart card
`programs in Hong Kong (one) and the United States
`(five). I chose these two regions because Hong Kong
`has one of the most highly successful e-purse programs,
`the Octopus card, and the United States has imple-
`mented a number of e-purse programs, some of which
`have been more widely adopted than others. I find
`that the most successful among these programs tend
`to have the following characteristics: a captive audi-
`ence that drives critical mass, such as those found in
`the transportation industry or government sector; an
`affordable cost structure relative to other payment in-
`struments; compelling incentives to consumers and
`merchants; and a technology that is well tested and
`addresses standards issues before the rollout.
`
`Carol L. Clark is a payments research manager at the
`Federal Reserve Bank of Chicago. The author gratefully
`acknowledges the assistance of Erin Davis, Juan A. De
`Jesus, David Doyle, Tamara Kidder, Graham Mackenzie,
`John Scaggs, Barbara Straw, Eric Tai, and Joey Wong in
`the completion of this study and the helpful comments on
`previous drafts by Sujit Chakravorti, Geoffrey Gerdes,
`Richard Porter, Tara Rice, and Leo Van Hove.
`
`34
`
`4Q/2005, Economic Perspectives
`
`GOOG-1036
`Google LLC v. RFCyber Corp. / Page 1 of 18
`
`
`
`Below, I survey the theoretical framework of pre-
`vious smart card studies, provide an overview of the
`payments environment in Hong Kong and the United
`States, and analyze six e-purse programs in these two
`regions and the factors that contributed to their success
`or failure. Then, I discuss the implications of my find-
`ings for future e-purse programs.
`
`Literature review
`One of the greatest challenges in the adoption of
`a new payment device is establishing a critical mass
`of users. Regardless of the type of technology used,
`consumers are reluctant to use a new payment instru-
`ment if few merchants accept it, and merchants will
`refuse to accept the device because the cost of install-
`ing and maintaining the supporting technology infra-
`structure, like card readers, may be prohibitive, unless
`enough consumers want to use it. New payment mecha-
`nisms gain momentum when enough people use them,
`which leads to widespread acceptance by the merchant
`community. Critical mass, however, is not only relat-
`ed to the number of users but also to the actual levels
`of usage because the program’s profitability is generally
`dependent on high transaction volumes (Goldfinger,
`1998). As Rochet and Tirole (2003) observe, merchants
`cannot benefit much from consumers that hold a pay-
`ment card but use it only sporadically. The more fre-
`quently the card is used, the more valuable it becomes
`to consumers and merchants. Therefore, frequent use
`is one of the keys to a successful e-purse program.
`Goldfinger (1998) estimates that a critical mass
`of one million users was needed for a smart card pro-
`gram to attain profitability due to the large fixed costs
`of the infrastructure, although these costs have likely
`fallen in recent years.2 To achieve this, Goldfinger ar-
`gues that program promoters have to be able to or-
`chestrate a large-scale deployment and initiate a
`migration/switching process from the existing pay-
`ment system to the smart card system. He takes the
`view that the benefits that smart cards provide cannot
`be fully realized if there is an alternative payment in-
`frastructure present. While this is certainly not the
`case for mature payment infrastructures—cash, checks,
`debit cards, and credit cards coexist at most retailers—
`there may be some validity to this argument in the
`case of an emerging payment instrument like an
`e-purse.
`In another study, Van Hove (2004) examines data
`on 16 e-purse systems in Europe. Van Hove finds that
`successful programs are in countries that are relatively
`small geographically or have phased introductions;
`that have online debit card systems that are fairly
`popular or cannot be used for low-value payments;
`
`that have stakeholders that quickly agree on a com-
`mon solution so there are no incompatibility problems;
`that have major banks committed to and participating
`in the program; and that have support from key players
`that operate and support one or more of the following:
`public telephones, parking meters, vending machines,
`or public transportation.
`Chakravorti (2004) finds three other necessary
`conditions for a viable new payment instrument: There
`must be benefits that are not provided in existing pay-
`ment instruments for at least certain transactions; con-
`sumers and merchants must be convinced of these
`benefits and, possibly, provided with incentives to
`change their behavior; and the new system must be
`perceived as secure, with adequate measures against
`credit risk and fraud.
`
`Payments environment in Hong Kong and
`the United States
`As I explained in the introduction, I am interested
`in comparing programs in Hong Kong and the United
`States because Hong Kong has one of the most highly
`successful e-purse programs, the Octopus card, and
`the United States has implemented a number of e-purse
`programs with varying degrees of success. As figure 1
`shows, Dove Consulting (2003) reported that in 2003
`electronic payments surpassed other types of payments
`for in-store purchases for the first time in the United
`States. However, cash was still the most popular pay-
`ment vehicle.
`Cash is used even more widely in Hong Kong.
`Eric Tai, chief executive officer of Octopus Cards
`Ltd., indicates that Hong Kong residents use coins
`
`FIGURE 1
`How U.S. consumers made in-store
`payments, 2003
`
`Check
`15%
`
`Cash
`32%
`
`Debit card
`31%
`
`Credit card
`21%
`
`Other
`2%
`
`Source: Dove Consulting (2003).
`
`Federal Reserve Bank of Chicago
`
`35
`
`GOOG-1036
`Google LLC v. RFCyber Corp. / Page 2 of 18
`
`
`
`and currency 50 percent of the time. Checks are used
`for retail transactions, where credit and debit cards are
`not accepted, and credit cards have become increasingly
`popular, with over nine million in circulation in 2001
`(Bank for International Settlements, Committee on
`Payment and Settlement Systems, 2003). Interestingly,
`however, in Hong Kong, e-purse transactions are now
`growing faster than either debit or credit card trans-
`actions—Euromonitor International (2004) reports
`that they increased by 8 percent in 2003, compared
`with 7 percent growth in debit card transactions and
`2 percent growth in credit card transactions.
`While the United States and Hong Kong have each
`implemented a number of e-purse programs, only Hong
`Kong’s Octopus card, which began in the niche trans-
`portation industry and extended outward to retailers,
`has been widely adopted by consumers and a diverse
`number of merchants. More than 95 percent of Hong
`Kong’s residents aged 15–65 carry the card. Over
`50,000 smart card readers accept Octopus at public
`transportation terminals, convenience stores, fast food
`chains, leisure facilities, parking meters and garages,
`pay phones, personal care stores, photo booths, photo-
`copiers, school snack shops, supermarkets, taxis, and
`vending machines (Tai, 2005). In August 2005, Octopus
`announced an apparel retailer will accept the card at
`its Hong Kong locations. Some e-purse programs in
`the United States that began in niche markets are cur-
`rently successful, but on a much smaller scale.
`Octopus processes over nine million transactions
`each day with an average daily transaction value of
`about HK$65 million (US$8.3 million) amounting to
`about 2 percent of Hong Kong’s gross domestic product
`(GDP) in 2003 (U.S. Department of State, Bureau of
`East Asian and Pacific Affairs, 2004).3 Retail purchases in
`Hong Kong using the Octopus card grew from 5 per-
`cent in January 2002 (Trintech Group Plc, 2003) to
`17 percent of total transactions in August 2005 (Wong,
`2005). With about US$1.4 million in average daily re-
`tail transactions, Octopus takes in more in a single day
`than the widely reported Mondex and Visa Cash trial in
`New York City did during the entire 15-month program.4
`As I mentioned earlier, most payment analysts
`agree that smart card adoption in the United States has
`been slower than in the rest of the world because the
`United States has an advanced telecommunications
`infrastructure that can verify magnetic stripe credit
`and debit card transactions quickly and cheaply online.
`This results in relatively low fraud levels and relatively
`high levels of satisfaction among businesses and con-
`sumers with the current systems. Smart card applica-
`tions may offer more value in other parts of the world
`
`with less highly developed telecommunications in-
`frastructures and higher fraud incidences.
`The business case for smart cards in the United
`States also depends on a number of other factors. There
`are issues related to who would pay for the extra chip
`on the card and to what fees merchants would pay on
`a per transaction basis. In Hong Kong, merchants ap-
`pear to be paying lower rates on Octopus transactions
`than on credit card transactions.
`There are also differences in the technology used
`for stored-value cards in the two regions. Octopus pro-
`vides e-purse capabilities on a contactless smart card,
`which means the card does not have to be inserted into
`a card reader like credit or debit cards. Instead, it is
`held close to the reader and payment is registered in
`0.3 seconds. Meanwhile, Duetto cards offered by the
`coffee chain company Starbucks, payroll cards that are
`used instead of direct deposit or paychecks by some
`firms to deliver an employee’s pay, and gift cards of-
`fered by various retailers in the United States provide
`stored-value capabilities on magnetic stripe cards. There
`are two ostensible reasons for using magnetic stripe
`cards rather than contactless smart cards in the United
`States: the cost of equipping stores with chip reading
`terminals and the desire to include Visa, MasterCard,
`or private label branding since these cards are processed
`by online readers.5 In addition, some payment providers
`in the United States offer contactless smart cards but
`link purchases to credit card or debit card accounts
`rather than to an e-purse—examples include Exxon-
`Mobil’s SpeedPass, Bank of America’s QuickWave,
`and MasterCard’s PayPass. In an interesting develop-
`ment, in December 2004, the Washington Metropolitan
`Area Transit Authority began piloting 20,000 Master-
`Card branded magnetic stripe cards that also contain
`a stored-value chip for transportation (Garback, 2005).
`
`Case studies
`I examine six e-purse case studies that began in
`“closed-loop” environments in Hong Kong and the
`United States, meaning they were offered to what one
`might call a captive audience, such as one found in a
`military facility or university campus. The e-purse pro-
`grams that were tested in open-loop environments in
`these two regions have failed outright, such as the
`Mondex and Visa Cash trial in New York City cited
`previously.6 I chose the case studies to represent a cross
`section of industries that have implemented e-purse
`programs in recent years: transportation, government,
`and higher education. The Octopus card’s e-purse trans-
`action volumes and values are among the highest in the
`world. The Ohio Electronic Benefit Transfer program,
`which has higher transaction values and volumes than
`
`36
`
`4Q/2005, Economic Perspectives
`
`GOOG-1036
`Google LLC v. RFCyber Corp. / Page 3 of 18
`
`
`
`Octopus, is the largest smart card program for adminis-
`tering food stamps in the United States. The Univer-
`sity of Michigan Mcard represents one of the largest
`university deployments of an e-purse in the United
`States. The University of Central Florida UCF Card
`is one of the few campus e-purse programs still in
`operation. The Navy Cash™ card and the EagleCash
`card programs are two of three smart card programs
`administered by the U.S. Department of the Treasury
`for the U.S. Armed Forces. A synopsis of the six pro-
`grams, as well as a detailed discussion on how each
`card works, is included in the appendix.
`
`Octopus card
`The Octopus card began in the niche transporta-
`tion industry when Hong Kong’s five leading compa-
`nies for trains, buses, ferries, and subways formed a
`joint venture in 1994 to oversee the implementation
`of a smart card system. After three years of develop-
`ment and trials, they launched Octopus in 1997. In
`1999, 7-Eleven stores in Hong Kong became the first
`locations outside the mass transit system where riders
`could add value to cards. The convenience store chain
`liked the speed and ease of the contactless technolo-
`gy so much that it installed readers in its stores in the
`following year so that consumers could pay for goods
`using Octopus. In time, consumers began to press other
`retailers to accept the card as well (Ramstad, 2004).
`A number of factors were crucial to the success
`of the Octopus card: the support of five transportation
`companies; the interoperability of the system; the man-
`ner in which critical mass was established by lever-
`aging the captive and niche transportation industry;
`the reliable technology; and the compelling incentives
`offered to consumers and merchants.
`Factors influencing success
`Octopus has the support of Hong Kong’s five ma-
`jor transportation companies. Although some of these
`companies compete directly for riders, the savings they
`achieved by implementing a shared smart card system
`appear to have outweighed any competition concerns
`(Poon and Chau, 2001). This also implies that the prof-
`it-sharing scheme the transportation companies worked
`out is equitable enough to induce cooperation. For con-
`sumers, the development of a single interoperable system
`means they can access any public transportation in Hong
`Kong with the same card. In contrast, 40 miles from
`Hong Kong in Macau, two bus companies launched sepa-
`rate incompatible e-ticket systems that failed to reach
`critical mass because traveling in the area typically
`requires a combination of buses and most people were
`not willing to carry two different cards (Uzureau, 2003).
`
`Octopus has also been free of technology-related
`problems, unlike several smart card programs that
`have had trials in the United States. Very few failures
`of the Octopus card were reported during the first
`month of operation. On average, station personnel
`needed to resolve problems in only one out of every
`11,000 journeys (Wynne, 1998).7
`Octopus also uses radio frequency identification
`(RFID) technology, which allows commuters to wave
`their card (or a purse or wallet containing the card) within
`4 inches of the reader at the ticket barrier to register
`payment within 0.3 seconds (BusinessWorld Publish-
`ing Corporation, 2002). Thus, an Octopus card trans-
`action takes less time than a cash transaction in which
`one may have to wait for change, and takes significant-
`ly less time than the typical credit or debit card trans-
`action in which magnetic stripe technology is used.
`Moreover, the durable smart cards have a potential
`life span of about 100,000 transactions (Tai, 2005).
`And Octopus’s functionality has been embodied in a
`variety of forms, including key chains, mobile phones,
`and watches.
`What about incentives? Initially, Octopus offered
`consumers a 10 percent savings and a 100 percent sat-
`isfaction guarantee to increase adoption in the trans-
`portation sector and to remove uncertainty about the
`new technology (Tai, 2005). These incentives, along
`with the simplicity, speed, and convenience of the
`system’s technology, resulted in over three million
`cards being issued during the first three months and
`established a critical mass of smart card users who
`were familiar with RFID technology.
`Metro and rail transportation operators offer mul-
`tiple ride tickets on the Octopus card and single ride
`tickets on magnetic stripe cards (Wong, 2005). This
`is significant because over 70 percent of Hong Kong
`residents use some form of public transportation each
`day (Poon and Chau, 2001) and are more likely to use
`the multiple ride tickets offered by Octopus. Tai (2005)
`reports that constraining multiple ride tickets to Oc-
`topus cards elicited little consumer dissatisfaction. Metro
`and rail transportation operators provide discounts to
`Octopus cards over single ticket cards; the discounts
`vary according to the distance traveled. Smart card adop-
`tion for metro riders is 90 percent and for rail com-
`muters over 80 percent (Wong, 2005).
`Transportation operators for buses, minibuses, and
`ferries accept coins or Octopus cards, and fares are
`the same for each payment method. Octopus card adop-
`tion on these transportation lines is somewhat lower
`compared with the metro and rail lines—70 percent
`for ferry lines, about 80 percent for minibuses, and
`over 80 percent for buses. Although buses, minibuses,
`
`Federal Reserve Bank of Chicago
`
`37
`
`GOOG-1036
`Google LLC v. RFCyber Corp. / Page 4 of 18
`
`
`
`and ferries do not consistently offer a discount to
`Octopus cardholders as do the metro and rail lines, they
`do sometimes launch promotional campaigns that of-
`fer discounts to Octopus cardholders only.
`Once a critical mass of smart card users was es-
`tablished in the transportation industry, the proven
`technology was used to branch out into the retail mar-
`ket, where consumers were offered a number of ben-
`efits that helped foster adoption. Octopus is a single
`convenient, multipurpose card that speeds retail trans-
`actions and replaces cash for small purchases. In con-
`trast to other e-purse programs, Octopus actually allows
`cardholders to make purchases up to a negative value
`of HK$35 (US$4), so long as the card contains a posi-
`tive value of HK$0.01 before the purchase. Once the
`card has a negative value, it must be reloaded before
`it is used again. Octopus recovers the negative balance
`through the deposit and purchase price of the cards.
`For a detailed discussion of the types of Octopus cards,
`deposit amounts, and card costs, see the appendix.
`Merchants also enjoy a number of benefits.
`Octopus reduces cash handling and in-store queues,
`and increases customer loyalty by allowing merchants
`to offer ad hoc discounts to customers using the card.
`It is difficult to determine the cost to retailers of ac-
`cepting the card, since data on hardware costs and
`merchant fees are confidential. The World Bank’s web-
`site indicates that Octopus has a two-part transaction
`fee. There is a HK$0.02 charge for every transaction
`to cover the costs of technical support, computer op-
`erations, and replacement cards and a 0.75 percent
`charge on the transaction value to cover card-control
`operations, legal, marketing, and depreciation costs.
`Therefore, a HK$10 transaction would include a fee
`of HK$0.02 plus HK$0.075, or HK$0.095 (Rebelo,
`1999). However, Octopus Cards Ltd. has indicated
`that these transaction charges vary depending on mer-
`chant volume (Cheng, 2004).
`Despite the uncertainty about exact costs, it ap-
`pears likely that retailers in Hong Kong benefit from
`lower transaction fees for the Octopus card relative to
`transaction fees for credit cards, which vary from 2 per-
`cent to 4.5 percent (Morgan and Snee, 1997). Although
`new locations like McDonald’s are accepting Octopus
`(Tai, 2005), some merchants still find Octopus fees
`to be too expensive. In CardTechnology, Balaban
`(2005) reports that a few retailers like Starbucks have
`reduced the number of outlets that accept Octopus.
`
`Ohio EBT program
`In the United States, the U.S. Department of
`Agriculture, Food and Nutrition Service (FNS) has
`shifted qualified low-income families from paper
`
`food stamp coupons to electronic benefit transfer (EBT)
`cards. The EBT program was designed to reduce fraud,
`to eliminate the cumbersome manual processes asso-
`ciated with issuing and redeeming paper food stamps,
`and to lessen the stigma associated with being a tradi-
`tional food stamp recipient. In 2003, 9.1 million U.S.
`households redeemed an average of $1.7 billion in
`food stamps every month using EBT cards. To reduce
`fraud, the system creates an electronic record of each
`transaction that can help identify where food stamps
`are trafficked or exchanged illegally (U.S. Department
`of Agriculture, Food and Nutrition Service, 2004).
`States have taken different approaches to admin-
`istering the EBT program. Forty-eight states have im-
`plemented magnetic stripe systems that require online
`authorization from a host computer that keeps track
`of value. Two states, Ohio and Wyoming, use offline
`smart card systems that store value on a computer
`chip resident on the card. The state of Ohio has an-
`nounced, however, that it is discontinuing its smart
`card program, Direction Card, which has been in place
`since 1996, and is seeking bids for an online system
`(Welsh-Huggins, 2003).
`Factors influencing failure
`John Scaggs (2005), Ohio’s EBT project director,
`indicated that the decision to discontinue Ohio’s offline
`system was based on cost, as well as on the failure of
`credit card companies to build a smart card infrastruc-
`ture, which had been anticipated when the program
`was implemented in the mid-1990s. The online system
`will be installed no later than June 2006.
`The decision to discontinue the program followed
`a 2002 study by Abt Associates, Inc. (2002) that com-
`pared Ohio’s program with the findings of the three
`most recent EBT system evaluations. These included
`the online system in the state of Maryland; the offline
`pilot in Dayton, Ohio, on which the Direction Card
`system was later built; and the offline system in the
`state of Wyoming. The study found Direction Card
`was more expensive than Maryland’s online system,
`but less expensive than the offline systems in Dayton,
`Ohio, and Wyoming. Abt Associates estimated that
`the total operational costs of the Direction Card system
`were 56 percent higher than Maryland’s system due
`to more expensive hardware, software, and local agency
`costs. The Direction Card was 29 percent less expen-
`sive than the Dayton pilot because of the larger scale
`of the Direction Card program, the lower costs of
`building the Direction Card system upon the Dayton
`pilot, and the decreased technology costs resulting
`from technological developments that emerged after
`the Dayton pilot was deployed. The Direction Card
`
`38
`
`4Q/2005, Economic Perspectives
`
`GOOG-1036
`Google LLC v. RFCyber Corp. / Page 5 of 18
`
`
`
`was 43 percent less expensive than Wyoming’s pro-
`gram due to lower local, state, customer service, data
`center hardware, software, POS, and card costs (Abt
`Associates, Inc., 2002).8
`The Abt Associates study also reviewed advan-
`tages and disadvantages of the Direction Card from
`the perspective of retailers. Ohio’s merchants received
`free terminal installation, initial user training, and
`maintenance; however, they never found a cost-effec-
`tive way to integrate the offline system into their ex-
`isting online POS devices. Therefore, Ohio retailers
`bore the ongoing costs of training staff to use the sepa-
`rate terminal, not to mention lost counter space. There
`were also differences in costs related to equipping
`store lanes with EBT POS devices. Even though the
`number of POS devices given to large retailers by Ohio
`was more generous than the FNS mandated, large re-
`tailers did not have enough terminals to equip every
`checkout lane in the store. To do so, they would have
`to pay for extra POS devices. In contrast, online mer-
`chants that integrated EBT transactions into existing
`POS devices could service EBT customers in any lane.
`However, most online systems did not pay the costs
`of integrating cards into existing POS devices, which
`then shifted the costs of doing so to the merchant.
`Those online merchants that did not integrate EBT
`transactions into existing POS devices either used the
`state’s allotment of EBT terminals or paid for extra
`equipment, but online terminals were less expensive
`than offline terminals (Abt Associates, Inc., 2002).
`The Abt Associates study also compared the ex-
`periences of Ohio’s EBT cardholders with those of
`EBT recipients accessing online systems. While Ohio
`recipients had higher levels of service due to hands-on
`training at Direction Card system offices, this special-
`ized training also required extra time from the card-
`holder and sometimes necessitated an added trip to
`the local office. In contrast, states with online systems
`piggybacked on the widespread use of magnetic stripe
`cards for other applications, as well as typically pro-
`viding cards, training materials, and personal identi-
`fication numbers (PINs) by mail. There were also
`differences in loading value onto the card. Ohio EBT
`recipients are required to load their benefits at any one
`of three stores of their choosing or at their local food
`stamp office. In contrast, EBT recipients with magnet-
`ic stripe cards do not have to load value on the card
`at any specific location because value is stored on the
`central computer server (Abt Associates, Inc., 2002).
`In addition, EBT smart card recipients may have
`experienced more confusion about the current value
`on their cards. Almost 90 percent of the Direction
`
`Card calls to customer service centers were to check
`the card balance. Similarly, the majority of customer
`service calls for online systems were to check account
`balances. The report indicates, however, that there
`may be an additional reason for balance inquiry calls
`for offline cards. The Direction Card system deducts
`the purchase amount from the chip on the card at the
`time of purchase. Information about the transaction is
`sent to the central computer server via batch processing
`at the end of the day when the beneficiary’s account
`information is updated. In contrast, online systems veri-
`fy transactions real-time against the central computer
`server, and balances are updated immediately. Thus,
`offline card users may have been confused by the bal-
`ance information on the audio response unit, which
`obtains information from the central computer server
`on a lag basis, compared with their knowledge of the
`available balance based on known card transactions
`and expected benefits (Abt Associates, Inc., 2002).
`The Abt Associates report also considered the
`transferability and adaptability of Ohio’s Direction
`Card to food stamp programs in other states. The main
`obstacle cited was the need to build a system from
`scratch, since few retailers have POS devices capable
`of reading smart cards because consumer demand for
`these cards has not reached a critical mass. Moreover,
`EBT recipients outside Wyoming and Ohio benefit
`from the interoperability of online systems, allowing
`them to access benefits in 48 states. Ohio and Wyoming
`recipients can only use smart cards in their own respec-
`tive states, unless an out-of-state store is specially
`equipped to accept them.9
`
`Campus smart cards
`Numerous e-purse programs have been imple-
`mented in closed-loop college and university environ-
`ments for a variety of reasons: Students are open to
`new technologies; universities are able to implement
`more secure IDs that are not as easily duplicated as
`magnetic stripe cards; and schools are able to reduce
`administrative costs and to generate transaction fee
`income. Despite these benefits, most campus smart
`card trials have failed. In University Business maga-
`zine, Villano (2004) reports that of the approximately
`50 schools in the United States that implemented smart
`card programs from 1997 to 2002, only a handful are
`still using them and relatively few take full advantage
`of the capabilities the technology provides. Failures
`are attributed to the high costs of offline systems com-
`pared with those of online systems, lack of interoper-
`ability, and delays in batch processing that mean card
`balances may not be updated for up to 24 hours.
`
`Federal Reserve Bank of Chicago
`
`39
`
`GOOG-1036
`Google LLC v. RFCyber Corp. / Page 6 of 18
`
`
`
`University of Michigan
`One of the largest e-purse deployments at a cam-
`pus in the United States began at the University of
`Michigan in 1995 in response to students’ and merchants’
`requests that the school’s Entrée Plus system be ex-
`tended off campus (Mayer, 1996). Entrée Plus was a
`funds pool into which parents put money at the begin-
`ning of the year for meals at residence halls and snack
`bars, as well as for on-campus purchases at vending
`machines, bookstores, and laundry facilities (Mitchell,
`1998). As it turned out, it was not feasible to expand
`the Entrée Plus system off campus, so the university
`developed a proprietary offline POS network called
`the Mcard (Mayer, 1996).
`In June 2001, the university announced that smart
`cards would be gradually discontinued and replaced
`with magnetic stripe cards (Avisian, Inc., 2002). A
`number of factors influenced the failure of the offline
`system, including technology problems; the overall
`cost relative to online systems; the lack of a critical
`mass of users and merchants; confusion related to
`concurrent programs on the card;10 and the apparent
`lack of a business case at the outset.
`Factors influencing failure
`In terms of technical problems, outdoor card readers
`did not function well in cold weather, and transaction
`times took longer than cash (Mitchell and O’Brien,
`1999). In addition, the chip on the Mcard malfunctioned
`nearly one-quarter of the time when it was first im-
`plemented (Doyle, 2005). Some cashiers had not
`been properly trained, which resulted in delays, and
`students became frustrated and eventually mistrustful
`of the card (Michigan Daily, 1996).
`Over time, the school found that the Mcard was
`more useful as a coin substitute than as a paper cur-
`rency replacement. As such, the school required tech-
`nology that would support pay phone, parking meter,
`and central parking facility transactions, but these were
`beyond the capability of the system (Doyle, 2005).
`Faced with an obsolete technology, the university
`sought bids from Visa, Mondex, and Proton to update
`the program. However, costs were considered too
`high, and the school announced it would revert to an
`online system (Kuykendall, 2001).
`Moreover, usage of the Mcard was lower than
`expected.