The Fresno Drop

On a Thursday in September 1958, Bank of America put 60,000 credit cards nobody had requested into the mailboxes of Fresno. The network that came out of that mailing became Visa, and almost nobody who paid for it was around to see what it was worth.

№ 007March 20269 min readPedrum Shokouhi

Mailboxes at the head of Second Avenue in Olivehurst, Yuba County, California, photographed in February 1940.

Joseph Williams

Customer Services Research Department

The Fresno Bee gave it six paragraphs. On Thursday, September 18, 1958, the paper’s readers were following the Communist shelling of Quemoy and a plan to reorganize the city’s police and fire departments, and on an inside page, wedged between the business briefs and the livestock report, they could learn that Bank of America had a new credit program. The program was already at their houses. Some 60,000 BankAmericards landed in Fresno’s mailboxes, each carrying a line of $300 to $500, and nobody receiving one had applied for it. A small personal loan at a bank branch had long meant sitting down with an officer, reciting your family history, and coming back with your wife to sign a note; now the note had been skipped and the money was in the envelope. The man behind the mailing was Joseph Williams, who after the war in Germany had driven to San Francisco to ask A.P. Giannini for a job, and who now ran a small in-house think tank called the Customer Services Research Department. He had never been a loan man at Bank of America. He and his team had set the credit limits, in Joseph Nocera’s account, on nothing but intuition, and he fully expected people to take its arrival calmly. Asked about launch day years later, Williams remembered it without ceremony: “There wasn’t a big parade or anything.”

The mailing looks reckless until you see the problem it solved, which is still the first problem of every network anyone tries to start. A card is worthless without shops that take it, and no shop will pay for a card that nobody carries. Small banks had been trying cards since the 1940s, and Williams studied the failures and came away encouraged: those programs had been too small and badly built, and he saw no reason Bank of America’s should be either. His answer was to stop recruiting cardholders and manufacture them. Fresno was picked because 45 percent of its families already banked with Bank of America, and because it sat far enough from anywhere that a flop would stay local. A salesman could now tell a shopkeeper, as a plain fact, that 60,000 local people would hold the card on a given day, and more than 300 merchants signed before the launch, small ones first, since Sears and Montgomery Ward ran credit operations of their own and regarded the bank as a poacher. The price was 6 percent of every sale plus $25 a month for the imprinter. Kenneth Larkin, a Bank of America officer, remembered a druggist with three bookkeepers chasing balances of $4.58 and $12.82 who nearly knelt in gratitude.

Los Angeles

June 1959

A test that careful deserved a year and got about three months. Word reached the bank that another California lender was planning a card of its own, a friendly merchant slipped over a copy of the plans, and the caution evaporated. Cards went to Modesto and Bakersfield within three months, to San Francisco and Sacramento by March 1959, and to Los Angeles by June. By that October some 2 million cards were out, more than 20,000 merchants had signed, and the rival never launched at all. Then the bill came. Williams had assumed 4 percent of accounts would fall behind, the rate on the bank’s installment loans; the real figure was 22 percent, and he had not built a collections department because he had not expected to need one. In Los Angeles crooks learned to read the floor limit off a stolen card and make hundreds of small purchases beneath it, thieves took blank cards from a warehouse and offered to sell them back, and a list of customers who should never receive a card was, by some mix-up, used to send each of them one. When colleagues asked what he was doing about the red ink, Williams told them, “All is well.” He resigned in December 1959, fifteen months after Fresno, with the official loss at $8.8 million. Count the advertising and overhead, Nocera reckoned, and it was closer to $20 million.

Clark Beise

The Managing Committee

What happened next is the part I find hardest to explain, because the sensible move was to quit. Smaller banks had been quitting for years, and Chase Manhattan, which launched its own card five months after Fresno, would be gone by 1962. The facts put to the bank’s managing committee, chaired by its chief executive Clark Beise, were awful, and nobody could say whether the program could be saved. In the bank’s own telling, which Nocera treats as a legend reasonably close to the mark, the committee looked ahead, saw credit cards in the future of the country, and took comfort in the thought that “every conceivable mistake had already been made.” They stayed. Williams’s people were reassigned as far from credit cards as the bank could send them, and the business went to the installment-loan men, who built the collections department, set up a fraud unit, took cards away from people with no intention of paying, and cut the merchant fee to as low as 3 percent for some of the larger stores they needed. An open letter went out to 3 million California families, part apology and part sales pitch.

Only 233,585 cards were in use in 1960, against some 2 million sent out. The card turned its first profit in the spring of 1961, and the profit for that year was $179,000, a number nobody in those meetings had yet seen and one that would not have comforted them. At that rate the bank would have needed 49 years to earn back the official loss and more than a century to earn back the larger one.

The hole and the first shovels

BankAmericard results at Bank of America, in dollars of the day

Uni-Serve

1962 to 1965

Williams’s story wants to be a clean injustice, the pioneer bled white while latecomers feasted, and the bank refuses to play its part. The losses turned out to be the best protection it ever had. Rivals knew about the early disaster and knew nothing of the later profits, so from 1958 until 1966 Bank of America had the California card business to itself. Annual sales on the card rose from $59 million in 1960 to $400 million in 1968, and the profit reached $12.7 million. An institution can carry a loss on its books and wait. Williams had no balance sheet. Former executives told Nocera he was forced out, Williams insisted he was not, and either way, Nocera said, he became the fall guy. Yet even Williams complicates the tidy version, because in 1962 he formed a company called Uni-Serve to buy Chase’s struggling card operation, paying $9 million, the value of its unpaid billings, and taking on a second bank’s failure. American Express bought Uni-Serve in 1965 and he stayed on until 1968, a successful second act, the New York Times called it. Holders of that card were later sent BankAmericards or Visa cards, so the second card he tried to save ended up folded into the first one.

Dee Hock

Columbus, Ohio, 1968

The second half of the story belongs to a man with no use for credit cards. Dee Hock had taken a steady job at National Bank of Commerce in Seattle in the mid-1960s after sixteen restless years in finance companies, intending, as he later put it, to “retire in place,” and he and his wife had long since sworn off every debt but the mortgage. Then his bank licensed BankAmericard and made him run it. Bank of America had begun franchising the card across the country in 1966, with rival California banks organizing a card of their own, and the franchise repeated Fresno on a national scale, with banks mailing cards to any list they could find, fraud everywhere, and clearing so slow that paper piled up in back rooms. In 1968 the bank called its licensees to Columbus, Ohio, where the meeting dissolved into accusations. Hock stood up and suggested they study the problem properly, and the relieved room made him chairman of a committee and went home. By June 1970 he had talked Bank of America out of its own creation, and control of the system passed to National BankAmericard Inc., a nonstock, for-profit membership corporation owned through nontransferable rights of participation, with Hock as chief executive. Four months later, on October 26, Congress made Williams’s method illegal, allowing a card to be issued only “in response to a request or application.” By then banks had dropped 100 million cards on the country.

The card became Visa in 1976 and 1977, and Hock ran the system until May 1984, when he resigned at 55. Three months later, with a successor in place, he dropped out of sight, to a ranch in Pescadero, a move he afterward described as “trading money for time, position for liberty, and ego for contentment.” The organization he left had no shares for anyone to sell. Mitchell Waldrop wrote that wealth, along with authority and decision making, was pushed out to the members, and Visa’s own filings show that Visa U.S.A. had no stock authorized, issued or outstanding as late as September 2007.

Then, in March 2008, Visa went public at $44 a share, the largest American IPO to that date, priced at $17.9 billion and then enlarged, for net proceeds of $19.1 billion. JPMorgan Chase made at least $1.36 billion in the same week it agreed to buy Bear Stearns, Bank of America’s listed take was $675.3 million, and that November JPMorgan held 23.3 percent of the banks’ class B shares to Bank of America’s 12.4 percent. It is tempting to read that as a verdict, the heir of the bank that got out of cards in 1962 collecting twice what the heir of the bank that stayed in collected. The corporate records do not allow it. The company called Bank of America in 2008 was run from Charlotte and was legally the old NationsBank, which had absorbed the San Francisco bank in 1998, and JPMorgan was by then Chase plus J.P. Morgan plus Bank One. Neither stake can be traced cleanly back to 1958, and the institution that paid for Fresno had stopped existing in any form Clark Beise would have recognized before the largest check arrived. Visa’s net income for fiscal 2025 was $20.1 billion, about a thousand times the larger estimate of the hole, in unadjusted dollars.

A network starts with someone willing to absorb a loss nobody else will touch, and an institution that lasts can collect for a while, as Bank of America did from 1961. The people do not. By 1994, when Nocera wrote, virtually everyone involved in the decision to keep the card alive was dead, and the IPO was still fourteen years off. Even that may flatter the decision. Fresno is a plane that came home; the smaller banks that gave up on cards before it took their losses too, and nobody writes essays about them. What separated Bank of America from them may have been little more than its size, more than 700 branches and the means to survive being wrong for long enough to find out whether it was. I would ask any business paying one side of a market to show up who will still be around to collect, and whose name goes on the failure if there is nothing to collect.

Williams died at his home in Atlantis, Florida, on November 8, 2003, a little over four years before the IPO. He was 88. The New York Times called Fresno and the California mailings after it “the official dawn of the bank credit card,” and a few paragraphs later it gave the loss to the decimal.

---

Notes

1. Unless noted, the Fresno scene, the quotations, the in-house history and the card figures are from Joseph Nocera, “The Day the Credit Card Was Born,” Washington Post Magazine, November 4, 1994, adapted from his book A Piece of the Action (1994), which drew on the bank’s own account and on interviews with Williams and Kenneth Larkin.
2. The card count is 60,000 in Nocera and in the New York Times and Los Angeles Times obituaries of Williams. Britannica and several other references give 65,000. I have followed the sources closest to the bank’s records.
3. The count of earlier cards depends on what is counted. Nocera has about a dozen attempts at an all-purpose card by the mid-1950s; Lewis Mandell, in the Times obituary, has more than 100 banks offering plans by 1955; the Federal Reserve History essay “Electronic Point-of-Sale Payments” (2024), citing the Federal Reserve’s 1968 study of bank card plans, has almost 100 banks starting to issue cards in the 1950s. The Los Angeles Times obituary dates small banks’ local plans to the 1940s.
4. Population: Nocera gives “around 250,000.” The 1960 census counted 133,929 in Fresno city (U.S. Census Bureau, Working Paper 27, table 19) and 365,945 in the county (the Bureau’s Population of Counties by Decennial Census, compiled by Richard Forstall), so the essay leaves the figure out.
5. The official loss of $8.8 million and the estimate “closer to $20 million” are both Nocera’s. A bare $20 million appears in Britannica and in the 99% Invisible episode “The Fresno Drop,” which also gives delinquency as one in four; Nocera’s figure is 22 percent. The payback arithmetic is mine: $8.8 million and $20 million divided by $179,000, in dollars of the day. It is hindsight, since Nocera places the committee’s meetings before the clean-up that produced the profit, and gives them no date.
6. Nocera dates the first profit to April 1961 and Britannica to May 1961, hence “spring.”
7. Williams’s departure, the dispute over whether he was forced out, Uni-Serve and the “fall guy” judgment: Douglas Martin, “J.P. Williams, 88, Bank Card Creator, Dies,” New York Times, November 21, 2003; Myrna Oliver, “J.P. Williams, 88; Banker Was Instrumental in Making All-Purpose Credit Card a Viable Idea,” Los Angeles Times, November 24, 2003. Neither obituary says who owned Uni-Serve, what American Express paid or what Williams received.
8. Hock: M. Mitchell Waldrop, “The Trillion-Dollar Vision of Dee Hock,” Fast Company, October 1996; Evan Sparks, “Nine Young Bankers Who Changed America: Dee Hock,” ABA Banking Journal, June 26, 2017; Hock’s biography at deewhock.com. The 1976 and 1977 dating of the rename follows the Times obituary of Williams. None of these sources gives Hock’s pay.
9. The ban on unsolicited cards is 15 U.S.C. § 1642, added by Pub. L. 91-508, title V, § 502(a), October 26, 1970.
10. IPO: Katie Benner, “Visa IPO prices at record $17.9B,” Fortune, March 18, 2008; Tami Luhby, “JPMorgan Chase makes $1B-plus on Visa IPO,” CNNMoney, March 21, 2008, citing the prospectus and reporting that Visa sold additional shares. Net proceeds of $19.1 billion, holdings and the non-stock history are from Visa Inc., Form 10-K for fiscal 2008; the percentages are of class B shares only, as of November 2008. Fiscal 2025 net income is from Visa’s earnings release of October 28, 2025. The lineage is from Bank of America’s Form 10-K for 1998, in which NationsBank is the surviving corporation; a Chase Manhattan filing of 2000; and a Bank One Form 8-K of 2004.

P.S.I’m Pedrum Shokouhi. I’m obsessed with how companies get built, funded, and run. So I write about them, take apart their strategy like an investor, and build my own. This is where I share what I’m learning.