In what is arguably the first independent research of its kind, The Tower Group, based in Newton, MA, has recently published a study on the progress and pitfalls of smart card development worldwide. It includes a discussion of applications, analysis and case studies of smart card schemes in Europe.

Although Europe and the United States have different banking structures, American bankers can learn from the European experience, says analyst Cynthia Weaver, who wrote the report. Instead of focusing solely on the electronic purse, banks should also consider supporting debit/credit applications. "The business case for the electronic purse as a single application is very unclear," she says.

Tower plans more research on the subject and is contemplating an extensive study of the business case for smart cards in the United States. What follows are excerpts from "Smart Cards in Banking: The Future of Money?"

In recent years, there has been intense interest in smart cards, particularly as they apply to banking. Not only can smart cards store considerably more information than magnetic stripe cards, they can actually execute programs that are stored on the card itself. These features mean that smart cards lend themselves well to use in financial services. Their primary advantage is their ability to operate more securely and provide significant safeguards against card fraud.

While the vast majority of bank cards in circulation continue to rely on magnetic stripes, smart cards have caught on in a few countries, particularly in France. There are currently just under 50 million bank smart cards in circulation, over half of which are in France. France stands out as the first country and, for now, the only country to have placed microprocessors on virtually all credit and debit cards in circulation. While France was the pioneer for this technology and its application in banking, other countries have started to follow suit. In Germany, banks are in the process of placing microprocessors on all 55 million eurocheque (ec) cards in circulation. While Germany will become the world's smart card leader, as measured by the number of cards in circulation, the number of transactions initiated using smart cards will remain greater in France for the foreseeable future. The French are the world's most active bank card users, and Germans lag far behind. In Switzerland and Austria, banks are replacing all magnetic stripe ec-cards with smart cards. In Belgium, the Netherlands, Spain, Finland, Portugal, and Denmark smart bank card schemes are being or have been rolled out nationally.

These national rollouts will drive significant growth in the number of smart cards used in banking. We expect the number of smart bank cards to more than double in 1997 and to continue to show very strong growth into 1998. Over 95% of these smart cards will be in Europe. U.S. banks and credit card issuers have been slow to move to smart card technology. As a result, the U.S. is ten to 12 years behind Europe in its adoption of smart cards, and shows no sign of closing the gap. U.S. banks have only recently begun small and limited tests of smart cards. A nationwide rollout of smart bank cards in the U.S. is unlikely to occur for at least six or seven years.

Much of the discussion of smart bank cards has revolved around stored value or electronic purse applications. Stored value, as the name suggests, means that money is actually stored on the card, rather than centrally in the bank's accounting system. In effect, the card becomes the system of record and keeps track of its own balances. Indeed, in the English speaking world, the expressions "smart card" and "stored value" are used almost interchangeably.

In reality, smart cards and stored value cards have little to do with each other. Smart cards can be stored value cards, just as they can be regular credit, debit or ATM cards. Traditional magnetic stripe cards can also perform any of these functions. There is a difference between the underlying technology and the applications supported. The technology used to implement the card may be microprocessor-based (smart), magnetic stripe or, more rarely, optical read. The applications supported by a card, on the other hand, may be traditional credit, debit, ATM, or stored value.

stored value: unappealing to public

The use of stored value smart cards remains limited. Most bank smart cards in circulation are simply traditional debit and credit cards. While a new generation of smart cards that combines debit and stored value features on a single card is gaining momentum, we do not expect stored value payments to account for a substantial portion of transactions in the future.

Some schemes, such as Proton, Visa Cash, Mondex and Danmont are purely stored value, i.e., they are electronic purses and do not support debit or credit functions. In stored value schemes, the customer must, in effect, pre-pay any purchases by loading value on the smart card. In practice this is done at an ATM or, occasionally, at specially equipped telephones. In the case of Proton and Mondex, money is loaded onto the smart card by transferring it from the customer's current account on to the card. Danmont, however, operates differently. The cards are sold in fixed denominations, e.g., Kr 200 or Kr 100, and are sold in return for cash at banks or other outlets. Once they are used up, the cards are thrown away or recycled.

The Geldkarte, issued by German banks, is a multi-application card. In addition to the normal immediate debit function that ec-cards have long offered, Geldkarte has an electronic purse built in for small value transactions.

The largest smart card scheme in the world, Cartes Bancaires (CB), does not support stored value applications. Geldkarte, the system that is about to become the world's largest, does support this application in addition to traditional immediate debit. So far, stored value has not been an overwhelming success, mainly because customers derive little benefit from its use. Its only convenience lies in the shedding of a few seconds of transaction time that would normally be needed to enter a PIN.

One of the main advantages for merchants is the elimination of the need for on-line authorizations, thereby reducing telecom costs and time involved in transaction processing. Stored value schemes are also free of local blacklists, which can take up several megabytes of storage space on the local device. In the CB scheme, debit and credit transactions are authorized off-line, which forces each terminal to maintain a blacklist of cards that has been reported lost or stolen. The blacklist is updated daily. Geldkarte follows the same approach. In both cases, the blacklist update takes place automatically, and is therefore not a burden to the merchant. The size of the blacklist may become a problem, but as the cost of memory goes down, the few megabytes of memory necessary to store such a list is hardly likely to be a major impediment. By and large, merchants are fairly indifferent as to whether they receive stored value payments, or whether they receive their money from a debit card. In both cases the merchant is risk-free, as long as requisite precautionary steps are followed.

Clearly, banks are great proponents of the idea of electronic purses; they stand to make money on the "float," or interest-free loan from the consumer. Some smart card schemes do not reimburse the consumer in the event that the card is lost, even if the balances on the card are not spent. In this case, the issuing bank is in effect obtaining indefinite float on the money on the card. Not all stored value schemes operate this way; some will reimburse any unspent money on the card once the expiration date is reached. However, in either case, the banks receive extended, interest-free credits. While the value stored on each smart card will typically be small, the total in the entire system could be very large indeed. In addition, depending on the smart card scheme, banks usually do not have to process each individual transaction. Supporters of stored value schemes argue that large volumes of small value debit transactions (under U.S.$5.00) would be infeasible due to bank systems overload.

(In France), payments with a CB debit or credit card typically do not require on-line authorizations, and are therefore every bit as convenient as stored value from the customer's perspective. Furthermore, CB debit and credit cards are used for small value transactions, at payphones, metro tickets and motorway tolls. The cards also give the customer added advantages. For example, the consumer does not have to reload the card periodically and does not have to keep track of card balances, as he does when using a stored value function. More significantly, the consumer is less concerned about card theft or loss since CB cards are PIN-protected and can be blocked. Stored value cards cannot be blocked; when they are lost or stolen, anyone can use the balance remaining on the card. To the rational consumer, stored value cards are not very appealing.

The majority of smart card transactions are simply traditional debit and credit payments. This is because CB cards, which make up the bulk of smart card transactions, do not support stored value. We do not expect stored value transactions to become the dominant form of smart card payment in the foreseeable future. On the contrary, we expect that stored value applications will not be particularly successful. Consumers are likely to reject the notion, as it offers them only limited advantages and serious disadvantages. For reasons outlined above, the rational consumer would, given the choice, avoid stored value transactions in favor of deferred debit or credit. Germany will provide the first real test of multi- application smart cards that support both stored value and debit applications.

Clearly, CB and Geldkarte are the world leaders, followed by Proton, SEMP and Quick. CB, Geldkarte and Quick developed their own technologies, and have not pushed product expansion internationally by licensing their technologies to others. Proton, SEMP, and Danmont, however, have encouraged other entities to license and modify their systems, with hopes of becoming the international provider of smart bank card services. Increasing international travel, the impending European Monetary Union and the future of the Internet as a payment system have spurred these companies to establish themselves as flexible, something-for-everyone organizations. Mondex, too, has spent millions marketing itself as an international system, but has required that all member banks agree on any changes to the system. This has made licensing of the system by different countries nearly impossible, as the scheme lacks the flexibility to be made compatible with existing infrastructures.

Will Mondex thrive?

It is worth noting that each successful scheme is backed by the majority of banks in the country, the national payment network, and in most cases, telecommunications companies as well. Projects that are less important tend to have the support of only a few major banks.

Despite considerable marketing and public relations expenditure, the Mondex smart card scheme has not achieved anywhere near the market penetration of the other smart card schemes discussed in this Research Note. Other schemes can boast thousands of times more users than Mondex, and tens of thousands more transactions. While Mondex is clearly the technical peer of any of the other smart card schemes, it does not have the unanimous support of domestic banks. Some of the largest banks in the U.K. chose not to participate in Mondex, including Barclays and Lloyds. This dramatically increases the risk for banks that choose to introduce smart cards by themselves.

Mondex was originally conceived by NatWest Bank in the U.K. in 1990. Midland Bank joined as a partner in 1993, taking a 50% stake in the venture. Unlike other stored value smart card schemes, Mondex behaves more or less like cash. Rather than require clearing through banks, Mondex value is transferred offline, directly from one card to another, without any need for third-party clearing. Other stored value approaches, such as Geldkarte or Proton, require every transaction to be processed through the participating banks' data centers. Both Geldkarte and Proton keep track of the value remaining on cards, by maintaining "shadow accounts," which force reconciliation between balances stored on the card and accounts maintained on the banks' system. This approach allows banks using these schemes to reimburse customers who have lost their cards (or at least to return unspent balances when the card expires). With Mondex, if the merchant wishes to deposit money collected over the course of the day, an on-line transaction must take place. However, unlike other schemes, Mondex requires only a value total, not every single transaction with counterparty information. The merchant may also choose to spend the Mondex money directly, and never deposit it in an account.

Despite its unique features, Mondex has not been a success. Even though Mondex started trials in 1992, well before Geldkarte or Proton, it has only a limited number of cards in circulation. Highly publicized trials in Swindon, a town in southern England, have been in progress since mid- 1995, and have had little success. Since then, only about 10,000 users, or 5% of the population, have signed up for Mondex. To make matters worse, Mondex card holders have not been particularly active, and we estimate that less than 0.5% of payments in Swindon are now made using Mondex cards. This is a disappointment, especially in light of heavy investments made by NatWest and Midland to ensure the success of this trial. The low penetration rate cannot be explained by a lack of merchants accepting Mondex cards. In fact, 70% of retailers in Swindon agreed to accept Mondex payments when NatWest and Midland agreed to provide the necessary smart card readers at a much-reduced price.

Other limited trials have been started in Canada, Hong Kong and New Zealand. Chase Manhattan is scheduled to begin a trial of Mondex in New York in late 1997. There is little reason to believe that these trials will be any more successful than the trial in Swindon. Banks in continental Europe, by far the world's largest and most sophisticated market for smart bank cards, have reacted to Mondex with a profound lack of interest.

Barring a major breakthrough, we expect that Mondex will be forced to retreat from its pretensions of becoming the worldwide standard for smart bank cards within the next year or two. In the future, Mondex may find a place as a provider of smart cards for special purposes, such as universities or corporate cafeterias, but Mondex, as we know it, will not become the world standard for chip-based card payments.

In November, 1996, MasterCard agreed to purchase 51% of Mondex, with the intention of integrating Mondex technology into its own chip project. MasterCard will probably use the Mondex operating system to support its own electronic purse application, formerly MasterCard Cash. However, it seems that the "just-like-cash" function, for which Mondex is known, may be lost in the process.

Smart Card rise in europe

The number of smart cards used in banking is growing at a rapid pace. In 1997, the number of smart bank cards will exceed 100 million. European countries have shown the greatest enthusiasm for smart bank cards, accounting for over 95% of the global total. Smart cards seem poised to become the dominant technology for bank cards in Europe, and we expect that within three to four years the majority of bank cards in Europe will be smart.

However, stored value is unlikely to become the dominant application for smart cards. Because of consumer resistance to stored value, traditional debit and credit will continue to make up the bulk of transactions. In a certain sense, smart cards are the future of money, but rather than replacing cash with electronically stored value, smart cards will simply make traditional debit transactions more secure and more convenient.

In the U.S., things look decidedly different. There have been only a handful of very limited trials of smart cards and these have not been met with unqualified success. While the business case for smart cards to replace magnetic stripe cards is fairly strong, due to the smart card's potential to reduce fraud, U.S. banks have been slow to adopt the technology. The highly fragmented nature of the U.S. banking industry makes cooperation between banks difficult. Tight cooperation and coordination between banks is of key importance when attempting a national rollout of smart cards. In all of the European countries that have rolled out smart card schemes, the effort has been spearheaded by a consortium that encompasses virtually all of the domestic banks.

The lack of cooperation between U.S. banks, and the limited experience that the U.S. has using smart cards in areas outside of banking, such as health care or telecom, mean that adoption will be slow. These factors indicate that a large scale deployment of smart cards in banking is at least five to seven years away in the U.S.

Cynthia Weaver is an analyst for The Tower Group, based in Newton, MA. The firm can be reached at www.; Weaver can be contacted at cweaver

Subscribe Now

Access to authoritative analysis and perspective and our data-driven report series.

14-Day Free Trial

No credit card required. Complete access to articles, breaking news and industry data.