2ndOpinion.FYI中文
genius.wiki

#505 1970 · Dee Hock / National BankAmericard Inc. · Banking / payments infrastructure

Rival banks would never let a competitor run their shared card network, so Dee Hock designed a company nobody could own enough of to control it

the problem

Competing banks need to share one payment network but none will hand a rival control of it

background

Bank of America launched BankAmericard in 1958 and began licensing it to other banks in 1966, but licensing didn't solve the underlying conflict: every bank that issued the card was depending on infrastructure a direct competitor still owned and could change the rules of at will. The system itself was buckling under its own growth — authorization meant a merchant phoning a call center and waiting, interbank settlement was still paper tickets mailed between banks, and fraud and backlogs piled up faster than clerks could clear them.

The two conventional fixes for shared infrastructure both failed here: let one participant run it (unacceptable — no bank would trust a rival with the switch that set its own costs and access), or hire an outside for-profit vendor to run it (unacceptable for the same reason, one layer removed — someone would still own the chokepoint and eventually price it like one). Dee Hock, then a mid-level executive at a Seattle bank, was asked to chair the committee of BankAmericard licensee banks tasked with fixing the mess, and concluded the problem the committee had been handed was the wrong one.

what everyone would do

The available options were to let one bank run the network, since Bank of America had invented it, or contract a neutral for-profit vendor to operate it as a service — and both hand somebody an ownership stake in infrastructure that determines every other bank's cost of doing business, which is exactly what a rival bank will never accept from a competitor or trust from a profit-seeking outsider either.

what they saw

Hock treated the problem as one of ownership design, not operations: the reason banks couldn't cooperate wasn't a lack of goodwill or a technology gap, it was that ordinary corporate ownership always lets someone eventually accumulate enough of a stake to control the whole. If control simply could not be bought — no stock existed to buy — the trust problem dissolved without needing anyone to behave better.

the move

In June 1970 Hock persuaded Bank of America to give up central control and incorporated National BankAmericard Inc. as a non-stock membership corporation. Ownership was expressed not as tradeable shares but as non-transferable rights of participation, sized to each bank's own transaction volume, with governance rules set collectively by the member banks that used the network — no outside investor and no single member could acquire a controlling stake, because there was no stock to accumulate.

why it works

Making participation rights non-transferable and sized to each bank's own transaction volume meant no member's influence could ever outgrow its actual use of the network, and pooling governance among the members who depended on the system made them collectively their own regulator rather than subjects of one. Because no outside party or dominant member could capture the network, rival banks were willing to route their own transaction volume and cede rule-setting to a shared body — the very trust the two obvious ownership models could never produce, since both left a single owner in the room.

the payoff

Competing banks that would not have accepted a rival's ownership joined a network none of them individually owned, and the membership scaled globally through the 1970s: the network launched its BASE electronic authorization and settlement system in 1973, cutting what had been a phone call and a wait down to seconds, and NBI was renamed Visa in 1976 as its structure was extended internationally through IBANCO. Hock served as the network's first CEO until his retirement in 1984.

where it breaks

The design needs the shared asset to be a genuine coordination good every participant needs on the same terms — it does nothing for infrastructure where participants' interests diverge sharply on how it should be run. It also depends on a usage metric that itself resists gaming, since a corruptible proxy for participation just moves the capture problem rather than removing it. And the structure trades away access to outside capital: exactly this cooperative eventually converted to a normal stock company once its growth needed more money than member banks alone could supply, showing the model has a ceiling, not just a floor.

what came after

The non-stock cooperative structure survived essentially intact for nearly four decades and underpinned Visa's growth into the world's dominant card network before the company converted to a conventional public stock corporation in its 2008 IPO, once its capital needs outgrew what a member-funded cooperative could raise. Hock spent his later career writing and lecturing on what he called 'chaordic' organizations — the term he coined for exactly this blend of decentralized ownership and collectively enforced order.

references

  1. [1]Nine Young Bankers Who Changed America: Dee HockABA Banking Journal (American Bankers Association), 2017bankingjournal.aba.com
  2. [2]Electronic Point-of-Sale PaymentsFederal Reserve History (Federal Reserve System), 2024federalreservehistory.org

keep it

same kind of clever