American Express and French meal vouchers both run on the three-party model, different from the four-party model you met in episode 1. Here a single player issues the card and funds the merchant itself.
In the classic model there are four main parties: the customer and their bank on one side, the merchant and their bank on the other. They are connected by a scheme that sets the rules and carries the information, without ever holding the money. In the three-party model, the two banks, issuer and acquirer, are replaced by a single entity. Four parties become three.
The AMEX case: three parties, two contracts, one price.
Interchange and scheme fees therefore have no reason to exist. The merchant sees a single rate and an invoice with one line, the commission.
A four-party scheme recruits almost nobody itself: its members do it for it, hundreds of banks signing cardholders and merchants in parallel, each for their own book. And interchange moves money from the merchant side to the cardholder side without the scheme paying for it. A three-party model has neither members nor interchange: with no merchants, nobody wants the card; with no cardholders, no merchant has a reason to accept it. Both ends have to be held at once, and funded in-house.
Hence the sales pitch, the same at every three-party scheme: AMEX does not sell volume to the merchant, it sells a customer base, positioning itself as a business introducer rather than a transaction carrier. That model has a price, and it is visible in the accounts.
FY2025: AMEX takes in almost twice what Visa does, and keeps half as much.
Pure three-party schemes have almost died out worldwide, yet one still runs at scale in a G7 country. Dematerialised French meal vouchers run on the three-party model, and the French competition authority labels them that way in writing. Conecs, created in 2012 and owned in equal shares by Edenred, Pluxee, Swile and Up, carries card transactions outside the banking schemes. Conecs does not authorise payments and does not fund merchants: Conecs plays a role similar to CB and Visa.
Each Conecs member signs both sides: the company that buys the vouchers for its staff, and the merchant that accepts them. The scheme itself is shared: four competing issuers run on the same infrastructure, closed and independent from the banking schemes. For the restaurant owner that means four issuers, four contracts and four rates.
Meal vouchers are not a market curiosity. In France they are worth around €9.4 billion in face value, 5.4 million employees and close to 234,000 affiliated merchants. But Conecs is not accepted everywhere, and building acceptance merchant by merchant is expensive. It is the chicken and egg of concept 2, in a French version.
Meal-voucher cards now carry two schemes: Conecs and a banking scheme. The issuer buys acceptance from the banking scheme, which works everywhere immediately, and keeps Conecs wherever the merchant accepts it directly. For the merchant the difference is not neutral: when the transaction runs on the banking scheme, he pays two commissions on the same transaction, his acquirer's, interchange and scheme fees included, and the voucher issuer's. Going through Conecs therefore costs him less, despite a rate that looks high.
Still a payments concept that escapes you?
Suggest the next topic