Finance people only

3 More Than Money

“Perhaps you would like to see what our credit cards are like,” [Doctor Leete asks of his guest, Julian West]. “You observe . . . that this card is used for a certain number of dollars. . . . The value of what I procure on this card is checked off by the clerk, who pricks out of these tiers of squares the price of what I order.” [Mr. West, from Boston, has awakened 113 years in the future, in the year 2000.]

—Edward Bellamy, Looking Backward, 2000–1887 (1888)

The simple payment card has been around since at least the beginning of the last century. Hotels, oil companies, and department stores issued cards before World War I. In response to customer requests, Sears began offering lines of credit in 1910 to customers of “unquestionable respon- sibility,” although the Sears card came more than a decade later. Some large retailers gave cards to their wealthier customers that identified them as having a charge account with the store. By the 1920s, several depart- ment stores allowed cardholders to pay off their bills in monthly install- ments. Metal “charge-plates” with embossed consumer information were introduced by department stores in 1928. During the 1920s as well, oil companies issued “courtesy cards” for charging gas. By the end of World War II, charge cards were no longer a novelty, but they were about as far from the cards of today as barter was from coin.

Dining on the Cuff

Restaurants did not issue cards. In 1949, Frank McNamara, the presi- dent of a New York credit company, was having lunch in Manhattan. A year later, as we mentioned earlier, he had a thriving business based on

this experience. He was written up in Newsweek two years later: “Halfway through his coffee, McNamara made a familiar, embarrassing discovery; he had left his wallet at home. By the time his wife arrived and the tab had been settled, McNamara was deep in thought. Result: the ‘Diner’s Club,’ one of the fastest-growing service organizations.” (This is the earliest rendition of the story we’ve found. One journalist several years later gave the credit to Alfred Bloomingdale, then the president of Diners Club, and changed the meal to dinner.)

Following McNamara’s epiphany, people began carrying charge cards in their wallets. McNamara and an associate, Ralph Schneider, started small. Beginning with $1.5 million of start-up capital, they signed up fourteen New York City restaurants and gave cards away to selected people. By the card’s first anniversary there were 42,000 cardholders, each paying $18 a year for membership in the “club.” And 330 U.S. restaurants, hotels, and nightclubs accepted these cards; they paid an average of 7 percent of the cardholder’s bill to Diners Club. In March 1951, Diners Club handled $3 million of exchanges between cardhold- ers and merchants, and reportedly made almost $60,000 in pretax profit. At that pace, it was handling $35.5 million in transactions annually. Unlike store cards, Diners Club cards provided a broader medium of exchange—one that extended to at least all the merchants in the club.

And that club expanded rapidly. In 1956, it had an annual transac- tion volume of more than $290 million. The card was accepted at nine thousand establishments, according to the New York Times, from Anchorage to Tahiti. By then its merchant coverage had expanded beyond restaurants to auto rental agencies and gift shops—almost the gamut of travel and entertainment locations. Two years later, Diners Club had an annual charge volume of more than $465 million, and earned gross profits of $40 million from merchant discounts and cardholder fees.

McNamara and Schneider had not only discovered the idea of a general-purpose payment card; they had also discovered a pricing strategy that got both merchants and cardholders on board, and thus has been followed by payment card systems since. By 1957, Diners Club had raised the cardholder fee to $26, but had left the merchant fee at 7 percent. It nevertheless continued to earn most of its revenues, about 70

54 Chapter 3

percent, from merchants. Similarly, American Express currently earns over 82 percent of its revenues from merchants, excluding finance charges.

Diners Club faced competition soon after its entry. Information is spotty on some—National Credit Card, Inc., for instance, started its card program in 1951, operated in forty-two states, but had filed for bank- ruptcy by 1954. A 1955 Newsweek article referred to Trip-Charge with eighty-five thousand cardholders and nine thousand merchants after a year in business; it asserted that the founder, Sidney J. Rudolph, “is now a hairbreadth from realizing his hopeful company motto: ‘Charge Every- thing Everywhere.’ ” Esquire and Duncan Hines both had travel and entertainment cards, which merged in 1957. Gourmet magazine had a club for diners, too.

Merchant coalitions were another source of competition. Hotel owners balked at the fee they had to pay on charge cards. In 1956, the American Hotel Association established the Universal Travelcard. It didn’t charge participating hotels and rental car agencies any fee but billed cardholders the same $26 fee as Diners Club. The National Restau- rant Association, with sixty thousand members, signed on. One might wonder how they could get by without the 70 percent of the pie from merchant fees that Diners Club received. Part of the answer is that they didn’t do central billing. Each hotel billed its customers directly, although the card association did ensure payment. There was a railroad card and an airline card as well.

Some banks had also entered the payment card business by the 1950s, but their cards—sometimes called “shopper” cards—targeted a different cardholder and merchant base. These bankcards were typically held by “housewives,” as the newspapers of the day put it, and could be used only at retail stores in the locality the bank did business. Franklin National Bank started one of the first shopper cards on Long Island in 1951, and one hundred or so banks—mainly small, suburban banks in the Northeast—followed. While one hundred may seem like a large number, one should keep in mind that in 1951, there were 13,455 com- mercial banks as well as many more credit unions and savings and loans. Banks generally charged retailers the then-standard 5 to 7 percent mer- chant fee. Cardholders reportedly didn’t pay any direct fees and, as with

More Than Money 55

the other card programs, were supposed to pay their monthly bill in full. These cards were thus charge cards, though the press of the day called them credit cards.

Of the early major entrants, only Diners Club survived the decade as a stand-alone company. It bought Trip-Charge in 1956 and Esquire’s card program in 1958. The bankcards failed mainly because they had trouble signing up merchants, and only twenty-seven of the shopper cards were still in existence by 1957. The Universal Travelcard and the Gourmet Magazine Club card were swallowed in 1958 by a new com- petitor (American Express) that appeared during what became a critical year for the future history of the card industry.

Diners Club also expanded overseas in the mid-1950s, using franchise agreements to extend its reach to Europe. As in the United States, European hotel trade associations posed strong resistance to the travel and entertainment (T&E) cards, going so far as to expel members who accepted payment cards that required a merchant discount. Some hotels in England and Switzerland chose to flout the prohibition; they accepted the T&E cards and formed their own association. They also went a step further: the newly formed hotel association created the BHR credit card in the 1950s, which evolved into the EuroCard in the mid-1960s.

1958

Though planning had started years earlier, several competitors rolled out new cards in 1958. In September, Bank of America started a credit card in California. In October, American Express launched its national charge card, and Hilton Hotels spun off its hotel card into Carte Blanche.

Bank of America California did not have restrictions on branch banking in 1958. With an economy larger than Japan’s, California was able to support several large banks. Bank of America was the largest, with over six hundred branches throughout the state. It had started as Bank of Italy in 1904, founded by one of the greats of U.S. banking, A. P. Giannini. By 1958, Bank of America was the largest bank in the United States.

Despite its size, though, Bank of America was cautious about offering a payment card, even though one of its small competitors, First National

56 Chapter 3

Bank of San Jose, had started a credit card in 1953. Bank of America considered introducing its own card in 1954, but initially decided that there wasn’t a good enough business case. After studying the emerging industry over the next few years, it decided to introduce a credit card in 1958. Creditworthy customers would receive cards with limits of either $1,500 or $2,600; prior authorization would be required for purchases over $260; and a revolving credit option was available for some card- holders. Revolving credit was the innovation that distinguished this card from existing charge cards.

The bank conducted a market test in Fresno, California, in fall 1958. Three hundred retailers signed up initially, and every Bank of America customer in the Fresno area received a card. According to one study, “This mass mailing of 60,000 cards had been William’s [the executive in charge of the effort] solution to the problem of how to convince retail- ers that enough individuals would possess a card to make their partici- pation in the program worthwhile. His solution worked, for during the next five months another eight hundred Fresno-area retailers joined the newly named ‘BankAmericard’ program.” Bank of America had planned to track the financial results of the card in Fresno before going statewide, but fear of competition persuaded it to accelerate the launch. It expanded throughout the state during the following year. By the end of 1959, twenty-five thousand merchants accepted the card and almost two million California households had one.

Things did not go well at first: fraud was rampant, the number of delinquent accounts was five times higher than expected, large retailers resisted joining, and echoing an old theme, “Public criticism came from those who viewed credit as a societal evil.” The program lost $45 million in 1960. The bank worked on collection problems and reduced the mer- chant fee to as low as 3 percent to entice retailers. Delinquencies declined and the merchant base increased to thirty-five thousand in 1962. The card turned its first operating profit in 1961.

American Express American Express started as an express mail company in 1850. Money was one of the things people wanted to move around the country, espe- cially after the post–Civil War expansion of the rail network created national markets. Of course, people also wanted their money delivered

More Than Money 57

safely. The U.S. Post Office developed the money order; American Express introduced a competing product. Both products were subject to theft, and neither was a good substitute for cash.

The travelers cheque, invented by an American Express employee, was a significant advance over the money order. The cheques came in multi- ple denominations just like cash and had the dual-signature system (sign when you obtain, and sign when you cash) that remains the major secu- rity device to this day. Initially, people could cash travelers cheques only at American Express offices; later, they could cash them directly at merchants. The major selling point for consumers was security: American Express guaranteed payment, but it also assumed responsibil- ity for lost or forged checks. The product has remained a highly prof- itable one for American Express for over a century, although its popularity has declined steadily over the last decade. The profits all came from individuals. A consumer purchasing $500 in travelers cheques, say, would pay American Express a fee in addition to the $500, and Ameri- can Express would continue to earn interest on the amount invested in the cheques until they were cashed. Stolen or misplaced cheques that were not cashed or replaced would be pure profit.

A hundred years after its formation in Buffalo, New York, American Express was the world’s largest travel agency and operated the world’s largest private mail service. Between its cheques and travel offices, it was profiting enormously from the boom in international travel following the end of the Second World War. The number of American Express travel agencies grew from fifty at the end of the war to nearly four hundred ten years later. The company sold approximately $6.5 billion worth of travelers cheques in 1951, and by the end of the decade claimed to control 70 percent of the U.S. travelers cheque business.

The Diners Club charge card was a new competitor for American Express. By 1953, American Express had begun planning its response. It considered buying Diners Club in 1956, but rejected the idea. The company finally entered the charge card industry on October 1, 1958, with 17,500 merchant locations and 250,000 cardholders. It achieved this scale quickly by buying the Gourmet Magazine Club card and the Universal Travelcard. Within seven months of launching its card opera- tions, American Express had over 600,000 cardholders. By late 1960, it had a charge volume of over $500 million and 750,000 cardholders.

58 Chapter 3

American Express adopted a slightly different pricing policy than Diners Club. It initially set its annual fee $1 higher (in 1958 dollars) than Diners Club’s $5, thereby suggesting that it was the more “exclusive” card (in 2002 dollars, the American Express annual fee was $31 while the Diners Club fee was $26). But it set the initial merchant discount slightly lower than Diners Club’s 7 percent: 5 to 7 percent for restau- rants, according to their sales volume; and 3 to 5 percent for the recal- citrant hotel industry, depending on the hotel guest’s charge level.

American Express struggled at first. Even a charge card involves extending credit for a time, and American Express, unlike the founders of Diners Club, had no experience doing this. By 1961, with losses mounting, it considered selling the business to Diners Club, but decided that such a sale might not pass muster with the Justice Department. Instead, American Express hired George Waters, later known as the “Father of the Card,” to run its card operations. Waters started putting pressure on customers who had not sent their payments in on time. And he raised the annual fee to $39, and later to $48. Despite the increased fees, the American Express payment card system continued to grow. By the end of 1962, there were 900,000 American Express cardholders who could use their cards at 82,000 merchant locations. In 1962, almost four years after its launch, the card operation posted its first (small) profit.

Today, American Express accounts for 14 percent of the dollars transacted on payment cards. By that measure, it is one-third the size of the Visa system (see figure 1.2 in chapter 1) and one-half the size of MasterCard. Diners Club, on the other hand, has shrunk to almost nothing within the United States. The charge card pioneer spread itself too thin in the 1960s and early 1970s in an attempt to counter the inroads made by American Express. In particular, Diners Club tried unsuccessfully to follow the American Express model by expanding into travel clubs and travel agencies, but it lost money on both endeavors and lost sight of the newly emerging competition from the bankcards.

Carte Blanche Hilton Hotels had issued a million charge cards for use at its worldwide hotel chain. After a failed attempt to buy Diners Club, it rolled its hotel card into a new general-purpose card company, the Hilton Credit Cor- poration, which distributed the Carte Blanche card. In 1958, it entered

More Than Money 59

with a low merchant fee, 4.5 percent, which soon dropped to 4 percent. The 600,000-member National Restaurant Association, which had been complaining about the 7 percent fees charged by Diners Club and American Express, threw its official support behind the card. Nonethe- less, as a result of issuing cards to the wrong people and an inefficient billing system, Carte Blanche became known in the trade as “Carte Rouge” for its steady losses.

What Happened to the Class of 1958 In 1960, a decade after the birth of the general-purpose payment card, there were three major national card systems. Diners Club, with 1.1 million cardholders, was still the biggest, but it faced competition from recent entrants American Express and Carte Blanche. These three, in turn, faced regional competition from BankAmericard in California and Chase Manhattan in New York City, though nothing else of significance. Chase Manhattan sold its card program to a subsidiary of American Express in 1962. This became the Uni-Card—a credit card that was available in the Northeast United States. It was sold back to Chase in 1969. Chase joined the BankAmericard association in 1972 and con- verted its cards to the BankAmericard brand.

Carte Blanche was sold to Citigroup in 1965. (Citigroup started as First National City Bank. In 1976, it became Citibank. Subsequently, the payment card operations were spun off into a separate subsidiary, Citi- group Global Consumer Group, which along with Citibank operates now under the Citigroup parent. For simplicity, we will refer to it as Citigroup regardless of the time period being discussed.) Pressured by an antitrust suit brought by the Justice Department, Citigroup sold Carte Blanche in 1968. The Justice Department was concerned that with Carte Blanche, Citigroup might limit development of the “Everything” credit card program it had started in 1967. When Citigroup quickly dropped the Everything card and had not introduced a replacement by 1978, however, the Justice Department relented, and Citigroup bought Carte Blanche back again. By that time, Carte Blanche’s share of credit card volume had declined to less than 1 percent.

American Express moved past Diners Club to become the industry’s volume leader in 1966. Diners Club continued to decline throughout the

60 Chapter 3

1960s, in part because it lacked the travel offices that American Express used to distribute its card during the industry’s early days. American Express also had a better T&E brand as a result of its travel offices and travelers cheques. Diners Club attempted to meet American Express head-on through travel and reservation system acquisitions, but it failed to make those profitable. Diners Club was sold to Citigroup in 1981. After the sale, Diners Club shifted its focus to affluent business travelers, trying to follow American Express’s successful upmarket strategy. Diners Club had 8.5 million merchant locations worldwide in 2002, including 2 million in the United States. That year, it had 0.5 percent of U.S. general-purpose payment card purchase volume and 0.7 percent of general-purpose credit and charge card purchase volume.

And the statewide BankAmericard became the worldwide Visa card.

The Birth of Co-opetition

Another watershed year for the emerging payment card industry was 1966, which marked the start of a battle between three competing busi- ness models for operating payment cards.

American Express, Carte Blanche, and Diners Club were mainly used for travel and entertainment, and thus became known as T&E cards. They did not offer credit beyond the time it took to get cardholders their monthly bills, which had to be paid in full. Nor was there a link to card- holders’ checking accounts. Many business travelers and wealthy house- holds had one of these cards, but most Americans didn’t. Data for 1966 are not available, but even by 1970 only 9.2 percent of households had one of the T&E cards.

Interstate banking regulations and other hurdles made it difficult for Bank of America to compete head-to-head with the three T&E cards. To take its card national, the California bank decided to franchise. In 1966, it announced that it would license its BankAmericard program to selected banks across the country. Each bank would operate the program inde- pendently using the BankAmericard name; merchants signed up by the franchisees would have to accept all BankAmericards, allowing con- sumers to use their BankAmericards at any participating merchant. Bank

More Than Money 61

of America charged the franchisees a royalty of up to 0.5 percent of card- holder volume and an entry fee of about $113,000.

Unlike T&E cards, bankcards did not charge cardholders membership fees, earning revenue from finance charges and merchant discounts instead. For example, the Chase Manhattan Charge Plan, introduced in 1958, charged cardholders 1 percent of the revolved (that is, unpaid) balance every month, while charging merchants 2 to 6 percent of sales depending on volume.

The BankAmericard franchise was not limited to the United States. Major banks in countries such as Canada, Columbia, Italy, Japan, Mexico, Portugal, Spain, the United Kingdom, and Venezuela signed up as inter- national BankAmericard franchisees around the same time as the domes- tic franchise system launch in 1966. In 1968, MasterCard also expanded internationally by forming alliances with EuroCard, the European card association mentioned earlier, and Banco Nacional in Mexico. The alliances allowed the MasterCard network and foreign networks to interoperate, but preserved each card as a distinct brand. In addition, MasterCard expanded in Asia by gaining member banks in Japan.

In the United States, within two months of the Bank of America fran- chising announcement, American Express, Carte Blanche, and Diners Club responded by offering their own franchise opportunities to banks. The American Express bankcard differed from the standard card: it offered a minimum $9,000 line of credit. American Express would split revenues with the banks: banks got a commission for signing up card- holders and revenues from credit provided by the card; and existing American Express cardholders could be converted to the bank program. American Express didn’t charge additional franchise or licensing fees. Carte Blanche priced its franchise at $45 for every $4.5 million in bank assets, with a $22,625 minimum fee. Diners Club offered to franchise its card for $22,625 to banks with less than $4.5 billion of assets and for $45,250 for banks with $4.5 billion or more. While the historical evi- dence is sketchy, it does not appear that the various efforts at franchis- ing these cards attracted any takers within the United States. (American Express also planned to franchise its Uni-Card credit card across the country in 1968. It had a million cardholders and eighteen thousand merchants in New England, New York, New Jersey, and Pennsylvania.

62 Chapter 3

Instead of following through on those plans, though, it sold the Uni-Card program back to its original owner, Chase Manhattan, in 1969, and Chase converted these cards to BankAmericards in 1972.)

For many banks, there were significant negatives to the franchise system. Major banks, including Wells Fargo in California and Chase Manhattan in New York, were not eager to sign up to issue someone else’s card. The successful franchise systems we’re familiar with— McDonald’s, the Athlete’s Foot, or Mail Boxes Etc.—typically involve a prominent brand name with outlets operated by unknown local entre- preneurs. Although some franchisees can become quite successful with multiple locations, they generally have little ability or desire to promote their brand name over the franchisor’s. This was not the case with the major banks.

Developing a proprietary card system was another option for banks. As we mentioned, Citigroup, in addition to owning Carte Blanche, had started its proprietary Everything card in 1967. Because Citigroup held a national banking charter and had customers across the country who were potential payment cardholders, it initially hoped to develop the Everything card into a national brand. Other banks found this option unattractive. While a national charter was, legally speaking, not neces- sary to issue credit cards around the country, some banks were reluctant to expand out-of-state.

Many banks found the answer in co-opetition. Banks competed for merchants and cardholders. Banks cooperated at the card system level by setting operational standards. Despite the dismal experience of the 1950s, many banks decided to start cards in the 1960s. They compared the problems of going it alone to the benefits of cooperation. They formed associations. Five banks in Illinois founded the Midwest Bank Card. By January 1967, nearly six hundred banks in Illinois, Indiana, and Michigan had joined; some of these issued one of the five original members’ cards. There were also two Michigan associations. Other banks across the country followed. Three New York City banks started the Eastern States Bankcard Association in June 1967. The state and local banking groups began to develop ties with other groups. The Inter- bank Card Association started in 1966. Early on it included banks in Buffalo, Pittsburgh, Milwaukee, Seattle, and Phoenix. At the same time,

More Than Money 63

several banks in California started the Western States Bankcard Association, issuing cards under the Master Charge service mark. By 1967, the California banks issuing Master Charge had joined the Inter- bank Association. By February 1968, Interbank had 286 banks in at least seven states.

It became apparent during 1968 that there were two competing national networks of banks: the BankAmericard franchise system, and the Interbank cooperative system. Banks—and groups of banks—started aligning with one or the other. “Just about every bank in the card field,” said Business Week, “is convinced that it must join one or the other network.” Bankers Trust in New York City went with BankAmericard and franchised the card to other banks in the New York area. Mean- while, Citigroup converted its Everything card to Master Charge and joined Interbank. Chemical and Manufacturers Hanover joined that association as well. For the most part, the larger banks had chosen Inter- bank over BankAmericard. In contrast to the BankAmericard franchise model, Interbank charged only a “modest” entrance fee and a small annual fee to cover the operating costs of the joint enterprise. And as noted, banks would be selling a brand they jointly owned, rather than that of another bank. This was an important point for banks that har- bored hopes of future national expansion when interstate banking restrictions were lifted—though in hindsight, that was still more than three decades away.

BankAmericard was not doing too badly by many measures. Under its franchise model, it had about 27 million cardholders and about 565,000 merchants by 1970, a sizable jump from 1.8 million cardholders and 61,000 merchants in 1966. But it was in the process of being overtaken by Interbank, which had attracted most major banks. And the fran- chisees were restless.

The franchisees quickly went from restless to rebellious. They had grown in importance to the system and wanted a voice in its future. In 1970, faced with this revolt as well as with operational problems, Bank of America agreed to convert the system into a membership-owned cor- poration: National BankAmericard, Inc. (NBI). NBI wasn’t an ordinary stock corporation; instead, its members had voting rights that they couldn’t buy or sell. Initially, NBI had 243 charter members, including

64 Chapter 3

Bank of America, Bankers Trust, and First Chicago as well as numerous smaller community and regional banks.

The brief period 1966–1970 turned out to be critical for the future development of the payment card industry. Three alternative business models battled against each other. The go-it-alone model had been the one used by American Express, Diners Club, and Carte Blanche for charge cards. Bank of America adopted this model in California, Citi- group tried it with its Everything credit card, and American Express tried it with its Uni-Card credit card. Bank of America and the three T&E card companies tried the franchise model. Finally, the co-opetitive model was used by Interbank and many other associations of banks across the country. The web of interstate and branch banking restrictions played a crucial role in the battle among these models.

By the end of the 1960s, the franchise model was dead in the United States. And the go-it-alone companies and the co-opetitives had gone in different directions. Sticking to what they knew best, the go-it-alones decided not to issue credit cards. American Express had ventured into credit with its Uni-Card, but decided to get out and did not try again for twenty years, as we will see. The co-opetitives, on the other hand, focused on issuing cards with a revolving line of credit. (Debit cards were soon added, but took until the early 1990s to have an impact.) Today, the co-opetitives account for 71 percent of card volume.

The basic idea behind co-opetition was clear as early as the Midwest Bank Card. The five founding members competed with each other for cardholders and merchants in the Chicago area. They cooperated in two related respects. They agreed to make their systems “interoperable.” A First National Bank of Chicago cardholder could use her card at every merchant who had signed up with any of the five banks. A Harris Trust Company merchant could accept as payment any card from these banks. What the banks lost in helping their competitors, they more than gained in making their own card more appealing to cardholders and merchants. Interoperability forced cooperation in another way. When a First National Bank of Chicago cardholder bought something at a merchant who was affiliated with Harris Trust Company, Harris Trust had to get reimbursed by First National. These banks faced the same problem as in the BankAmericard franchise: system cooperation was essential to

More Than Money 65

process the slips of merchant receipts that were growing exponentially with the number of participants in the system.

The two national associations encountered similar issues to Midwest Bank Card, only on a grander scale. MasterCard—which started out as the Interbank Card Association in 1966, changed its acceptance brand to Master Charge in 1969, and finally became known as MasterCard in 1979—and Visa—which began as BankAmericard in 1958, switched to National BankAmericard, Inc. in 1970, and settled on Visa in 1976— took cooperation further than any of the regional associations. First, they established rules and processes for settling transactions. Part of this involved how the merchant and the cardholder banks divvied up the transaction proceeds. Some of the regional cooperatives had initially exchanged at par so that the cardholder’s bank reimbursed the mer- chant’s bank for the entire transaction and the cardholder’s bank didn’t get any of the merchant fees. (The banks may have simply been apply- ing the check model of par exchange, which they soon found unsatis- factory for cards.) MasterCard and Visa both settled on an interchange fee—a percentage of each transaction that the merchant’s bank gave to the cardholder’s bank. (We discuss this further in chapter 6.)

Another area of cooperation was on the card brand. The banks that belonged to Interbank decided early on to use the Master Charge brand. That—and not the individual bank’s name—is what was most promi- nent on the cards from the late 1960s and early 1970s. Visa replaced BankAmericard as the brand name for that system in 1976. The co- opetitive felt that “standardizing the somewhat confusing array of blue-white-and-gold cards issued under different names in twenty-two countries around the world” would lead to greater acceptance of the card. Focusing on the system’s brand involved a trade-off basic to the co-opetitive model: choosing between doing things at the system versus the member level.

Overall, the co-opetitives chose to do most things at the member level. A few statistics are revealing. Association membership fees and dues for Visa and MasterCard, which cover the costs of the centralized opera- tions, account for only about 1.5 percent of the total direct card-related expenses incurred by members of these systems. Visa’s centralized activities are conducted by a staff of about 1,300 employees. To put this

66 Chapter 3

in perspective, these people comprise less than 0.5 percent of the total estimated employees involved in issuing Visa cards.

Regulation and Stagflation

American Express prospered during the 1970s. Between 1960 and 1977, real net income grew at an average annual rate of 16.6 percent. By 1977, American Express had eight million cardholders, bringing in $393 million in annual card fees. Its lead over Diners Club and Carte Blanche had widened dramatically. American Express had decided against offer- ing a credit card and had unloaded its Uni-Card credit card. About 200,000 of its cardholders had “corporate cards”—cards that employ- ers ask employees to use for expenses and that provide employers with detailed spending data. Sticking with charge cards, shunning credit cards, and focusing on corporate users was a profitable strategy for some time.

Meanwhile, American Express’s credit card competitors—the banks that issued credit cards and the two associations to which they belonged—struggled during this decade of government regulation, volatile interest rates, and economic stagnation. The economy went through several severe recessions in the 1970s and early 1980s (see table 3.1). The economy also experienced accelerating inflation, jumping

More Than Money 67

Table 3.1 U.S. recessions, 1969–1982

Peak monthly Duration unemployment

Start End (months) rate (%)

December 1969 November 1970 11 5.9 November 1973 March 1975 16 8.6 January 1980 July 1980 6 7.8 July 1981 November 1982 16 10.8

Note: The National Bureau of Economic Research defines a recession as “a period of significant decline in total output, income, employment, and trade.” Sources: National Bureau of Economic Research, U.S. Business Cycle Expan- sions and Contractions, <http://www.nber.org/cycles/> (accessed April 21, 2003); and U.S. Department of Labor, Bureau of Labor Statistics.

5 From Sardi’s to Saks.com

The [credit cards] are a new burden in our industry—a tax. And they don’t bring in more business—not in my experience.

—Philip Rosen, owner of New York City’s exclusive Café Chambord (1958)

“You’ll be the savior of my business,” [he exclaimed] when told of BankAmeri- card’s plan. The costs associated with managing his 4,500 [store] accounts were dragging him under.

—Drugstore owner in Fresno, California (1958)

Merchant acceptance of payment cards has expanded like ripples in a pond since McNamara cast the first stone. One day, a few brave mer- chants in a category take cards; soon most do. Over time, new categories of merchants take plastic. Some of the most recent are convenience stores, fast-food restaurants, and Web-based retailers.

Are cards accepted everywhere that you want to be? It only seems that way; there are many payments for which plastic still isn’t taken. Some taxis honor payment cards, but most don’t. Neither do most landlords or mortgage lenders. Nor does the babysitter. But past experience sug- gests that plastic will expand into these and other areas in which cash and checks remain the sole coin of the realm.

From Major’s Cabin Grill (the birthplace of Diners Club), which forked over 7 percent of the tab to Diners Club in the early 1950s, to 7- Eleven, which paid around 1.5 percent to Discover a half-century later, many merchants have believed that the benefits from taking plastic exceeded the costs. And part of the explanation for the expansion of cards is that merchants’ benefits have increased while merchants’ costs have fallen. Taking cards is more valuable because more customers have

them, clerks spend far less time processing transactions, and computer- ized billing and payment have improved. Merchant fees and other costs of taking plastic have plummeted.

Yet merchants complain about having to take a haircut every time a customer whips out plastic instead of a wad of cash. Indeed, organized merchant boycotts of payment cards have occurred repeatedly since hotels banded together in the mid-1950s to refuse to accept Diners Club, and they show no signs of abating. Most recently, various coalitions of retailers around the world have used the legal system to secure reduc- tions in what they pay for taking plastic. These are classic symptoms of a two-sided market—“Let the other side pay”—but merchant complaints have also led to a debate over whether there is too much plastic.

Getting to Be Everywhere You Want to Be

The spectrum of merchants who take plastic for payment has widened considerably over time. It began with upscale restaurants in 1950, and expanded through the decade to hotels, car rental agencies, and other travel-related businesses. American Express persuaded retailers with heavy patronage from travelers to take their cards, and in the 1970s it convinced upscale department stores such as Saks Fifth Avenue. American Express cards were attractive because the cardholders had the “businessman, traveler-tourist, and expense account profile” these merchants found appealing.

Later on, credit cards brought in many other retailers, from hardware to clothing stores, where more typical households shopped. Larger retail- ers were leery of credit cards because these cards competed with their profitable store cards. They started taking credit cards only when many households had one of these cards and wanted to be able to use it for payment. As J.C. Penney’s CEO explained, the nationwide department store chain finally decided to take Visa cards in 1979 because “we rec- ognized that we had fifteen million active accounts, while all Visa members have thirty-five million active accounts. . . . It seems logical to expect that a good number of people who don’t carry our card but who have Visa accounts would use Visa cards in our stores if they had the chance.”

116 Chapter 5

Plastic cards have spread from merchants who cater to the wealthy to those who cater to the poor. Even bail bondspeople take plastic these days. But cards have also spread to merchants who sell small-ticket items and to those with relatively low margins. Only about 5 percent of super- markets took plastic in 1991, while by 2003 almost all did. To make this happen, the co-opetitives and go-it-alone systems took steps to reduce supermarkets’ merchant discounts, and the EFT systems persuaded supermarkets to install PIN pads. The fraction of fast-food restaurants that took plastic cards has also increased; there were around 40,000 loca- tions (out of a total of around 120,000) that accepted payment cards in 2002, up from 24,000 in 2001. And in March 2004, McDonald’s announced plans to accept all major payment cards at its stores.

Web-based businesses are of course a natural for payment cards. Many of these dot.coms are virtual extensions of existing brick-and-mortar firms: Saks has Saks.com. Others, like Amazon.com and eBay, lead only a virtual existence. These Web merchants take payments almost exclu- sively through payment cards. The dollar volume of Web-based transac- tions increased at an annual rate of 54 percent between 1998 and 2001, and amounted to $30 billion in 2001.

In 1959, about 162,000 U.S. merchants took one or more payment cards. By 1971, this figure exceeded 820,000, implying an average annual growth rate of over 14 percent. By 2002, more than 5.3 million merchants took payment cards. This implies an average annual growth rate of about 6.2 percent since 1971.

Part of this growth has resulted from the increase in the number of merchants in the economy. If all restaurants take cards and the number of restaurants increases, there is an increase in the total number of mer- chants who take cards. It is useful to adjust the figures on merchant acceptance by the growth in merchants over time. For example, there were about 27 percent more merchants in 2002 than in 1971. When we adjust for the growth in merchants—that is, if we view growth in card acceptance as growth that is in addition to the growth in the total number of merchants—the average annual growth rate from 1971 to 2002 becomes a smaller but still impressive 5.4 percent.

As table 5.1 shows, by 2001, payment cards had become widely accepted in a broad range of merchant categories.

From Sardi’s to Saks.com 117

118 Chapter 5

Table 5.1 Plastic share of dollar volume by merchant category, 1996 and 2001

1996 (%) 2001 (%)

Total retail 31.8 46.2 Automotive 22.7 37.5 Food/grocery/drugstores 17.9 38.3 Drugstores 22.3 39.9 Grocery/supermarkets 17.2 38.7

Department stores 60.5 66.9 Specialty stores 41.8 55.9 Apparel stores 54.9 65.1 Hardware/home improvement stores 42.0 53.4

Travel and entertainment 42.1 47.9 Restaurants 25.0 35.9 Fast-food restaurants 1.8 5.3 Mid-priced restaurants 28.6 43.9 High-priced restaurants 51.0 60.8

Travel 68.9 74.7 Airline companies 84.5 88.2 Car rental agencies 86.8 89.6 Hotels/motels 80.0 80.7

Entertainment 19.1 25.0 Tickets 36.5 39.3 Movie theaters 3.4 12.9

Services/recurring expenditures 4.5 8.0 Insurance services 0.4 3.5 Utility/telephone 0.7 3.4 Phone company 1.0 5.5

Health care 17.5 25.1

Source: Visa U.S.A.

Getting Cheaper and Better

Merchants typically contract with an acquirer working on behalf of a card system (the acquirer may be the system itself in the case of the go- it-alones). The acquirer agrees to provide authorization services (using the card systems’ computer networks) for all cards carrying the logo of each system it represents and to reimburse the merchant within a set number of days for all charges that have been authorized. The merchant is typically guaranteed payment even if a cardholder never pays their bill or if the card is stolen—so long as the merchant follows the authoriza- tion procedures agreed to (such as comparing signatures on the slip and the card). The acquirer typically also provides related services to the mer- chant such as periodic billing information that can be integrated into the merchant’s accounts payable system.

The merchant pays a fee for these services—the merchant discount. In the case of the co-opetitives, this can be thought of roughly as the system’s interchange fee plus other costs that the acquirer has to cover. (These are sometimes shown separately on the merchant’s bill.) In addi- tion, the systems typically impose some other requirements on the mer- chant. Until recently, the merchant had to accept every card that carried the logo of any system with which the merchant had a contract. This provision—sometimes known as the “honor-all-cards rule”—prevented a merchant from taking cards only if the cardholder didn’t have cash, from taking cards from some kinds of customers but not others, and from picking and choosing among different kinds of cards offered by the system. Chapter 11 discusses a retailer lawsuit against the co-opetitives that resulted in an agreement by the co-opetitives to permit U.S. mer- chants, beginning in January 2002, to accept a system’s credit cards without having to accept its signature debit cards.

Another card system rule prohibits merchants from imposing sur- charges on customers who use payment cards. Merchants may give dis- counts for cash or checks if they wish, but they may not add separate charges for using plastic. This is known as the “no-surcharge rule.” Card systems appear to have imposed both these restrictions on merchants since the early days of the industry.

From Sardi’s to Saks.com 119

The Basic Economics of Taking Plastic You have finally realized your lifelong ambition of opening a hardware store in Minot, North Dakota. In addition to figuring out which brands of sledgehammers and duct tape to stock, you have to decide how you are going to let people pay and then arrange to implement each of the payment mechanisms you agree to accept. Should you take checks in addition to cash? Which, if any, brands of payment cards should you take? You cannot answer these questions without considering what forms of payment your customers would like to use, what forms your competitors are accepting, the costs of taking various forms of payment, and other factors.

Let’s assume that like most retailers, you have decided to take cash. What about checks? You might get more business by taking checks, but you have to worry about bounced or fraudulent checks. Bad check costs amounted to $6.14 billion in the United States in 2001. To address this, you might pay for a check verification service or restrict check accep- tance to customers you know or those who are in-state. (Some large retailers with a lot of check use, like supermarkets, set up their own check programs requiring customers to apply for check-writing privileges.)

Next, you consider whether to accept payment cards. As with checks, you may get some additional sales if you take plastic. Some of your cus- tomers may not be carrying cash or their checkbooks when they stop by your store, or they may avoid your store because they like to pay with plastic and your competitors take it. So when considering benefits, you need to take into account the volume of these “incremental sales” as well as the profits you will earn on them.

Your additional sales depend on the extent to which your customers have payment cards. Like J.C. Penney, many merchants started taking cards only when enough of their customers carried them. Even today, you probably won’t take a JCB card because the Japanese tourists who typically carry this card seldom frequent Minot hardware stores. Your incremental sales also depend on how your customers use their cards. People are more likely to want to use plastic for large transactions because they are less likely to have sufficient cash on hand, and they may be particularly likely to use credit cards if they want to finance the pur- chase. Cash is commonly used in small transactions but not for large

120 Chapter 5

From Sardi’s to Saks.com 121

ones, while the opposite is true for payment cards and checks, as figure 5.1 shows. If you plan on stocking a wide selection of expensive tools and gadgets in addition to nuts and bolts, you would expect that many of your customers would like to use their cards.

The value of the additional sales you get from accepting payment cards depends on how much you mark up the goods and services you sell. You will care much more about making an additional sale if you make a $20 profit on it than if you make a $2 profit. Across major retail categories, the average margin (measured as the difference between revenues and the cost of goods sold) in 2000 varied from 12 percent for car dealers to 44 percent for clothing and accessories stores. Hardware stores have an average retail margin of 34 percent, so you can pick up some signif- icant extra profits if taking cards increases sales.

There are other benefits as well. Most acquirers provide your store with convenient billing information. Cards, unlike checks, are

0%

10%

20%

30%

40%

50%

60%

70%

80%

Other

PIN debit

Signature debit

Credit

Checks

Cash

$500+$100–500$80–100$60–80$40–60$20–40$10–20$5–10

Figure 5.1 Payment methods by transaction size, 2001 Source: Visa U.S.A.

guaranteed not to bounce, and you don’t have to worry about pilferage or theft as you do with cash. Credit cards also enable your customers to finance their purchases and may thus encourage them to buy more. Historically, many larger stores found that it was profitable to provide customers with installment loans or store credit cards. But it was never efficient for a single hardware store to operate such a program. It comes as little surprise, then, that smaller stores generally took credit cards well before the major department stores.

Of course, you also have to consider the potential cost of accepting payment cards. There are fixed costs incurred in setting up payment card acceptance, including purchasing one or more card readers, setting up phone lines, and training staff. To accept PIN debit, you have to incur the additional cost of a PIN pad, which is not needed for signature debit, credit, or charge cards. These fixed costs don’t depend on your volume of payment card transactions, so you must have a large enough volume of such transactions to cover these costs. As well, you will incur variable costs—merchant discounts—when you take cards. These costs vary depending on the card. For a $100 purchase, they range from 41¢ for PIN debit cards to $2.64 for American Express (based on industry aver- ages; we don’t have available rates specific to hardware stores).

You also need to consider differences in processing transactions. Each payment mechanism takes time at the checkout counter—tying up your clerks and making your customers wait in line—and other processing costs. Both have proven hard to pin down, though numerous studies have tried. But as we discussed in chapter 4, based on casual experience and some evidence, cards seem to take slightly more time than cash and sig- nificantly less time than checks. Customers don’t like to wait in line, but having more checkout lines to make them happier increases costs.

Merchant Competition At this point, you are almost in a position to compare the value of taking one or more different kinds of plastic. (Many acquirers offer bundles of several cards, all of which work with a single card reader and will appear on a single statement.) But there is still one important element missing in the calculus: what your competitors are doing. If most of them have decided to let customers pay with plastic, then you will be at a compet-

122 Chapter 5

itive disadvantage if you don’t accept plastic as well. Perhaps by avoid- ing the cost of payment cards you could offset this disadvantage by offer- ing lower prices. In practice, as we saw above, plastic tends to either become almost universally accepted or ignored within a particular mer- chant category.

That is not a surprising result in competitive markets. Casual obser- vation suggests that merchants attempting to sell roughly the same goods to roughly the same customers tend to offer roughly the same amenities, all of which cost something to provide. Most restaurants have public toilets, restaurant decor reflects the prices on the menu, many stores have free customer parking, and your hardware store, like others, is probably open on Saturdays. Businesses that compete directly often have adver- tising campaigns that also compete directly, even if they don’t name each other. Almost every merchant incurs the costs of advertising and ameni- ties mainly in hopes of attracting customers from other stores. Of course, in the aggregate, merchants would be better off if none of them had to do any of this, but then shopping would be much less pleasant (or for some of us, even more unpleasant).

This sort of nonprice competition, which includes competing by accepting checks and payment cards, has a curious feature: it can be socially excessive. Consider advertising. If merchant A’s advertising simply shifts business from merchant B to A, the gain to A is likely to substantially outweigh the gain, if any, to society as a whole—that is, to all merchants, including B, and all consumers in aggregate. With a few more assumptions, one can argue that merchants as a group tend to advertise more than would be optimal from the standpoint of society as a whole. This “market failure” has been almost completely ignored in policy circles for one simple reason: even if there is too much advertis- ing in theory, it is not clear how in practice one could ever calculate the optimal amount of advertising, let alone attain it. One might tax adver- tising spending, for instance, but a tax that is too high would do more harm than good, and there is no way in practice to know how high is too high.

It is interesting that basically the same argument is taken seriously by some policy makers when it is made in the more complex context of mer- chant acceptance of plastic cards. The contention here is that merchant

From Sardi’s to Saks.com 123

A benefits much more than society as a whole by accepting MasterCard, for example, because by so doing merchant A steals business from mer- chants B, C, . . . and Z. Thus, merchant A will find it profitable to accept MasterCards even when the costs to society as a whole from that mer- chant doing so outweigh the benefits. And so, most likely, will merchants B, C, . . . and Z. Since the more merchants accept MasterCard, the more attractive MasterCards are to consumers, the end result could be too many transactions on MasterCards relative to the theoretical optimum. Moreover, so the simplest argument goes, since merchants are overeager to accept MasterCards, the MasterCard system will likely maximize its members’ profits by setting excessive interchange fees.

Of course, as in the case of advertising or merchant acceptance of checks, there is no way in practice to calculate what the ideal interchange fee or merchant discount should be. Furthermore, more complete eco- nomic models show that the profit-maximizing merchant discount may be below the socially optimal discount, so that reducing the discount may make society as a whole worse off.

For Merchants, Cards Have Gotten Better From the standpoint of the merchant, taking payment cards has become more valuable. More potential customers have these cards and want to use them for buying and financing more things than in the past. Out of every one hundred households, twenty-one had one or more payment cards in 1970, forty-five in 1983, and eighty-four in 2001. The number of customers with card-based credit lines has also increased: from sixteen in 1970, to forty-three in 1983, to seventy-three in 2001. All payment card systems have developed increasingly sophisticated computerized systems for authorizing and processing transactions, and acquirers have been able to offer better billing and accounting services to merchants over time.

The performance of payment cards relative to other payment media has also improved. The amount of time that it takes to complete a trans- action is one of the most important attributes of a payment system. There have been slight improvements in completing cash transactions over time—for example, cash registers now tell a clerk how much change to give. And there have also been some improvements in taking checks. But

124 Chapter 5

By contrast, multisided platforms—especially those in new markets— all too often require clean-sheet planning. With multiple yet interdepen- dent business constituencies to serve, costs offer little guidance for pricing strategies. By the same token, early entry may yield first-mover advantages or provide an instructive failure that simplifies the search for successful strategies by businesses that follow. And in light of constituent interdependence, changes in the business environment may have multi- sided effects that are difficult to anticipate.

Along with greater challenges go greater rewards to the nimble and the better capitalized. Many of the great companies of the modern era— think of eBay, American Express, Microsoft, and Cisco Systems—have prospered precisely because they have excelled at making multisided plat- forms work to their advantage.

Multisided Platforms and Price Setting in Payment Cards

McNamara gets credit for creating the modern payment card because he had the inspiration to turn the charge card from a single-sided product into a two-sided platform. Stores had allowed people to buy now and pay later since at least the early nineteenth century. Retailers started using charge cards in the early twentieth century as a convenient device for identifying and keeping track of customers with accounts. Installment loans and charge cards helped retailers sell things, but they were all part and parcel of a single-sided market—retailers selling goods and services to customers. These instruments may also have resulted in different customers paying different prices—for example, early on Sears charged credit customers more —depending on the terms of payment and finance charges.

McNamara turned the charge card into a two-sided platform by recognizing the opportunity to create a new payment device that could be used by many customers at many different merchants. Payment cards had been a vertically integrated activity engaged in by individual merchants. McNamara’s idea was to vertically disintegrate payment cards from the merchant and offer a program many of them could use. While this insight seems obvious in retrospect, it was not so at the time.

It Takes Two to Tango 149

The pricing model developed by McNamara and his partners formed the basis for the modern payment card industry. For transaction services, the merchant contributes most of the revenue and the cardholder receives the service for little or nothing after taking the float and other benefits into account. This point is worth emphasizing because of the far- reaching consequences it has for the business strategies followed by the payment card systems, but also because of its role in much of the litiga- tion that has enveloped the payment card industry during the current decade. Based on the available data, the share of American Express rev- enues coming from merchants has been between roughly 65 to 80 percent over the last four decades. And the share of MasterCard/Visa member revenues coming from merchants has been between roughly 60 to 75 percent during that same time. (Data are not available for all years. We should note that meaningful and consistent comparisons across time are difficult, as the share of revenues coming from merchants is affected by the mix of credit, charge, and debit volume, which has varied both across systems and time. Credit cards present an additional complication because certain revenues may be attributable to the financing rather than the payment function. We have excluded finance charge revenue from these calculations, assuming that they represent revenues associated with the financing function. Revenues from service fees, such as for late payment, are more difficult to apportion between the payment and financing functions, although such fees have not generally been at sig- nificant levels in the industry until recent years. For MasterCard and Visa, we have counted half of service fees as revenues from cardholders attributable to the payment function. For American Express, service fees are not reported on a sufficiently disaggregated basis to be incorporated in these calculations. Despite these difficulties in getting precise estimates, the numbers suggest broadly that as far as transaction services are con- cerned, the merchant has consistently provided the greater share of rev- enues to the system.)

Two-Sided Pricing Strategies by Go-It-Alone Systems The go-it-alone payment systems have two basic pricing instruments to manage the two-sided markets in which they compete. They can choose the merchant and cardholder fees to conduct the balancing act

150 Chapter 6

central to all multisided markets and to compete against overlapping or intersecting systems. A significant part of the competition in the payment card industry over the last fifty years has involved the strategic use of merchant discounts (and for the co-opetitive systems, interchange fees).

Diners Club entered with a zero cardholder fee, which it raised when it had gotten the critical mass of cardholders necessary to attract mer- chants. Various competitors in the early 1950s tried lower merchant discounts, but these did not generate a sufficiently larger number of merchants to get more cardholders, and without more cardholders, mer- chants didn’t have much incentive to take the cards, even with a lower discount.

American Express came in with a lower merchant discount and slightly higher customer fees than Diners Club. Bank of America, which started as a go-it-alone system, entered the California market with a merchant discount comparable to that of American Express, but it had a different business model. While American Express and Diners Club were target- ing travel and entertainment establishments patronized by high- spending and often well-heeled customers, Bank of America went after a much broader group of retailers where cardholders could use credit cards. Its merchant discount was 5 percent, while American Express had a sliding scale ranging from 3 to 7 percent.

When Discover entered, it offered merchant discounts that were sub- stantially lower than those charged by acquirers for the co-opetitives and by American Express. Initially, it had no merchant acceptance beyond Sears (which owned it) and had to persuade merchants to take another card brand. The lower merchant discount and the fact that the card was offered to twenty-five million creditworthy Sears cardholders helped Dis- cover attain merchant acceptance that exceeded that of American Express’s within eight years of its inception, and that today is almost equal to that of the co-opetitives.

The go-it-alones have also varied their merchant discounts to pene- trate additional merchant segments. American Express eventually decided that it needed to expand its merchant base to compete with the co-opetitives. It lowered its merchant discount for retail segments that had resisted accepting American Express, such as gasoline stations.

It Takes Two to Tango 151

Figure 6.1 shows how the average level of American Express’s merchant discount has varied over a forty-five-year period.

Interchange Fees and the Co-opetitives In the 1950s and early 1960s, most banks—even large New York City ones—could not create platforms large enough to harvest efficiently the indirect network externalities discussed above. Their geographic foot- prints were too small as a result of branch and interstate banking restric- tions. One exception was BankAmericard in California. The co-opetitive structure enabled banks to create platforms large enough to provide enough value through network effects to attract cardholders and merchants. These co-opetitives, however, which as we saw in chapter 3 emerged in many parts of the country in the mid-1960s, had two related business problems.

In these co-opetitives, banks pooled their cardholders and merchants together under a single brand. Thus, cardholders from one bank could use their cards at merchants signed up by another bank. To make this

152 Chapter 6

0%

1%

2%

3%

4%

5%

6%

7%

200219981994199019861982197819741970196619621958

Figure 6.1 American Express average discount rates, 1958–2002 Sources: American Express, Annual Report (various years); Bernstein Research, The Future of the Credit Card Industry; and various trade press.

work, the co-opetitives needed a set of rules. Who would bear the risk that cardholders wouldn’t pay their bills? Who would bear the costs incurred in serving merchants in that case? And who would bear the costs incurred by issuers? The answers to those questions affected the gains to trade that banks received from one another in the co-opetitives. Balancing the competing interests of banks was of primary importance in the early days of the co-opetitives, as we discussed in chapter 3.

The co-opetitives faced another problem—one that was far less appar- ent to them in the early days. The go-it-alone systems had two strings to pull to balance the demand of cardholders and merchants. Without more authority, the co-opetitives had no way to influence the relative prices paid by the two sides. The price level as well as structure would be what- ever resulted from the competition between banks for merchants, on the one hand, and cardholders, on the other.

The interchange fee emerged early on as part of the solution to the first problem and over time became the solution to the second one as well. Without an agreement on interchange fees, it is unclear a coopeti- tive card system could function successfully. The evolution of the inter- change fee at Visa and its predecessors is instructive.

Bank of America started as a go-it-alone system. It set cardholder fees and merchant discounts on its own. When it created a national franchise system, however, it needed to decide what to do when cardholders from one franchised issuer used their cards at a merchant affiliated with a dif- ferent franchised acquirer. It decided to have the acquirer pay the full merchant discount to the issuer. Acquirers didn’t earn any profits—or even cover their costs—on those transactions. Aside from reducing the incentives to acquire rather than issue, this arrangement obligated acquirers to report merchant discounts accurately to issuing banks. Issuers sometimes doubted that this was done, and the conflict and ill will created by this scheme was one of the reasons the Bank of America franchisees quickly became unhappy.

In addition to the mistrust between issuers and acquirers, the settle- ment process became more and more cumbersome. Initially, a merchant had to make a phone call to get authorization for large transactions. To save time, the merchant could accept transactions below a “floor limit,” commonly around $200, without calling. Each card transaction

It Takes Two to Tango 153

generated a card slip that had to go from the merchant to the acquirer, then to the issuer, and finally to the cardholder. Such a system was inef- ficient and conducive to fraud. Moreover, increases in the number of mer- chants, cardholders, and transactions bogged this “interchange” of transactions down in paper.

After the franchisees’ revolt, Visa (then NBI) instituted several changes that addressed the former franchisees’ issues. First, it set a formal inter- change fee that was uniform across members and thus not linked to the merchant discount fees charged by individual acquiring banks. This elim- inated any incentive to misreport those fees. The initial interchange fee Visa set in 1971 was 1.95 percent. The first formal methodology for setting the fee, developed by Arthur Andersen in 1973, identified two components of credit card operations: the financing function, and the payment function. (At that time, fees other than finance charges on credit cards were uncommon, both because most states prohibited such fees and because cardholders appeared unwilling to pay fees even in states that permitted them.) The payment function was referred to as the “merchant servicing” function because its primary beneficiary was assumed to be merchants. The methodology assumed revenues to the system would come either from finance charges or merchant fees. The ostensible purpose of the interchange fee was to reimburse issuers for costs that were not attributable to the financing function. So for example, the cost of funds for providing the grace period would be part of the interchange fee, while the cost of funds for providing revolving credit beyond the grace period would not. Issuers would be reimbursed for the merchant-servicing-related costs on transactions that were interchanged (that involved another bank). And as we discussed above, both Visa and MasterCard built computer systems that approved—or “authorized”— card purchases to speed up the process and reduce fraud. These systems automated both the authorization of transactions at the time of sale and the bookkeeping among members to settle transactions at the end of the day.

To complement the newly established interchange and authorization systems, the associations developed rules that basically concerned both who got what and who was responsible for what. For example, the asso- ciations negotiated rules for how to handle fraudulent charges or charges

154 Chapter 6

on which consumers default. As long as the merchants met certain terms, such as properly authorizing transactions and checking card numbers against a list of known fraudulent accounts (this process is automated now), the system guaranteed payment. If a transaction turned out to be fraudulent or a consumer refused to pay for it, the issuing bank was responsible for the charge.

The interchange methodology was revised substantially in 1981. By then, cardholder fees had become much more common, and it was believed that cardholders valued the payment function associated with cards rather than just the financing function. So rather than viewing the payment function as solely related to servicing merchants, this was now something of value to both sides. As before, costs were apportioned between the financing function and the payment function, but now issuers would share in paying for the costs of the payment function on interchanged transactions. In the framework of multisided markets, we see that the balance struck by Visa changed to accommodate changes in the value that cardholders and merchants placed on the system. When cardholders began to see more value in the platform, the methodology shifted to a more merchant-friendly approach. In addition to the new accounting scheme, Visa instituted lower interchange rates for merchants who moved from paper to electronic processing. All of these factors— the new interchange methodology, the new lower rates for electronic transactions, and as a consequence, more and more merchants switch- ing to electronic processing—combined to lower the average interchange fee for Visa throughout the 1980s, from an average of 1.7 percent in 1982 to 1.3 percent in 1989.

The interchange fee is a cost to acquirers that must be covered by the fees they charge merchants. As a result, the merchant discount for Visa and MasterCard includes the interchange fee. With intense competition among acquirers and thin margins in recent years in the United States, the interchange fee seems to be passed on to merchants. That is, changes in the interchange fee lend to changes in the merchant discount. (In the United States, large merchants typically pay an acquirer fee plus the inter- change fee; the effect on smaller merchants may not be as direct or imme- diate. Also, different business and institutional arrangements lead to different results in different countries.)

It Takes Two to Tango 155

Likewise, the interchange fee is a benefit to issuers; it is a source of revenues to offset their costs. Although it is harder to document, intense competition among issuers means that the interchange fee is largely passed on to cardholders in the form of lower fees and/or better features. Indeed, that is the reason why issuers can profitably provide low-fee or no-fee cards with various rewards even to individuals who do not revolve balances.

Over time, the co-opetitives had to keep three considerations in mind in setting their interchange fees. The first consideration is the role of acquirers and issuers as members of the association. This was not an issue when banks did both acquiring and issuing, so that interchange fees and expenses roughly balanced out for most members. It became more of a problem as banks specialized in either acquiring or issuing. And it became less of a problem as the association became more concentrated in the hands of banks that had delegated much of the arguing business to third-party firms.

Second, there is the concern about competition between the co-opeti- tives for members. This has become a particular issue recently as Mas- terCard and Visa have fought for near-exclusive arrangements with banks, and as membership duality has become less important. We discuss this in more detail in chapter 8.

Third, and foremost, is the effect of the interchange fee on cardholder versus merchant demand. To see this, consider what would happen to prices under different interchange fees given the intense competition in issuing and acquiring. In 1983, before the explosive growth of credit cards, the Visa interchange fee was 1.6 percent, the average merchant discount rate was 2.3 percent, the average annual cardholder fee was $16.86, and charge volume per account was $1,720. If the interchange fee fell to zero, the merchant discount rate would have fallen to 0.7 percent (assuming interchange fees were fully passed on to merchants)— a fifth of that being charged by American Express at the time—and the cardholder fee would have had to almost triple to $44.38 to keep the average issuer’s revenues constant. (Changes in finance or other charges might have occurred in addition to or instead of a change in the annual fee, and the overall level of card ownership and usage would likely have declined.)

156 Chapter 6