Fifty Years of Rules Nobody Wants to RewriteWide moat

Underneath the network sits a rulebook. Who is liable when a card is used fraudulently. How a disputed charge is reversed and on whose evidence. What a merchant may and may not surcharge. What happens to settlement when a member bank fails on a Friday afternoon. It runs to thousands of pages and it is the reason a shop in one country accepts a card issued in another without knowing anything whatsoever about the issuer.

The rules matter more than the technology because they allocate risk. A merchant accepts a Visa card not because the message routing is elegant but because it knows precisely what happens when the transaction goes wrong, and who pays. That guarantee cannot be shipped in version one of anything, which is why entrants tend to begin in the lowest-risk corner of commerce — small amounts, known counterparties, domestic only — and find it very hard to leave.

It is also why the network's moat is partly a liability-allocation moat rather than a technology one. The chargeback right is the single most valuable thing a cardholder has and almost never thinks about; the account-to-account rails that compete on price mostly do not offer it, and cannot easily add it without acquiring the underwriting function that makes it possible.

The same rulebook is Visa's principal legal exposure. Nearly every antitrust action against the company for forty years — the merchant litigation, the DOJ's interest in PIN-authenticated debit and the fixed acquirer network fee¹ — has been an argument about a rule rather than about a price.