Interchange: The Fee Structure That Shapes the Industry

When a merchant accepts 100 by card and receives roughly 98, the missing amount is not one fee but a stack, and each layer goes to a different party.

The total is the merchant discount rate, or merchant service charge. Pricing models differ, blended pricing quotes one rate, interchange-plus passes each layer through transparently, but underneath every model the stack is the same.

Gotcha: interchange is not what Visa or Mastercard earn. It flows through the network to the issuing bank. The network keeps only the much smaller scheme fees. Most public anger about card fees is aimed at the party that keeps the least of them.

Interchange is a balancing mechanism for a two-sided market, and taking that framing seriously explains behaviour that otherwise looks perverse.

A card scheme only works if both sides participate: cardholders will not carry cards no merchant accepts, and merchants will not accept cards nobody carries. But the costs fall unevenly. The issuer carries credit risk, fraud losses, the interest-free period, and the expense of acquiring and keeping cardholders. The merchant side captures much of the benefit: higher conversion, bigger baskets, guaranteed payment.

Interchange transfers money from the side that benefits to the side that bears cost, tuned by the schemes to maximise total participation. Set it too low and issuing becomes unprofitable, so banks stop pushing cards. Set it too high and merchants revolt, surcharge, or steer to other payment methods.

That is the theory the schemes advance, and regulators partly accept it while contesting the level. What nobody disputes is the mechanism: interchange is a deliberate subsidy flowing from acceptance to issuance, and everything downstream, from rewards to fee caps, is a fight over its size.

Follow the money on a premium rewards card and the loop closes neatly.

The cardholder earns, say, 1 percent cashback. The issuer funds that from the interchange it collects on the cardholder's spending. The interchange is paid by the acquirer, which prices it into the merchant discount rate. The merchant prices its costs into what it charges, spread across all customers however they pay.

So rewards are, to a first approximation, a transfer from everyone who shops to the people holding the richest cards, routed through the fee system. Card industry economists and central banks have made versions of this point for years; it is why premium cards carry higher interchange rates than standard ones, and why merchants dislike them specifically.

The design also explains issuer marketing. A bank that wins a big-spending rewards customer books an interchange annuity on that customer's entire future card spend. Sign-up bonuses worth hundreds are rational customer-acquisition cost against that stream.

Where regulation caps interchange hard, this engine loses fuel, and the observable result is exactly what the mechanism predicts: markets with capped interchange have visibly thinner rewards programmes.

The European Union regulated the number directly. The Interchange Fee Regulation, Regulation (EU) 2015/751, applied from December 2015, caps interchange on consumer card transactions.

Member States may set lower domestic caps, and several have. The regulation also bans schemes from making one card's acceptance conditional on another's, and requires interchange-plus transparency to be available to merchants on request.

The caps had exactly the effects the mechanism predicts. Merchant fees on consumer cards fell substantially. Rich cashback and rewards cards became rare in Europe compared to the United States, where consumer credit interchange remains unregulated and typically runs several times the EU cap. And issuers recovered revenue elsewhere: annual fees, currency conversion margins, and a decisive push toward premium commercial products, which brings us to the loophole.

The caps come with boundaries, and the most commercially important one fits in a reveal.

The other notable boundaries: the caps bind card-based transactions within the European Economic Area, with three-party schemes like American Express outside the core caps when they operate without licensing issuers, and cross-border transactions where a leg sits outside the EEA falling under separate, higher, negotiated rates.

For a merchant, the practical consequence is that the average card fee is a mix that depends on the customer base. A business hotel's card mix is expensive; a supermarket's is cheap. Two merchants with identical turnover can face materially different acceptance costs, and neither chose it.

The United States makes a clean natural experiment because it regulated one card type and left the other alone.

Debit interchange for large banks was capped under the Durbin Amendment to the Dodd-Frank Act, implemented by the Federal Reserve's Regulation II: a fixed cents-plus-basis-points formula that cut large-bank debit interchange roughly in half when it took effect in 2011. The rule has remained contested ever since, with a Federal Reserve proposal to lower it further and court challenges to its basis running through the 2020s, so treat the exact figure as a moving target and the structure as the durable fact.

Consumer credit interchange, by contrast, is set by the networks with no statutory cap, and typically runs in the range of one and a half to two and a half percent or more depending on card tier and merchant category.

The predictable results: the US has the world's richest credit card rewards, an enormous premium-card industry, persistent merchant litigation against the networks, and periodic legislative pushes to regulate credit interchange. Same mechanism as Europe, opposite policy choice, opposite equilibrium.

Within a market, interchange is not one number but a published grid. The networks maintain tables with hundreds of rates, and a specific transaction's rate is looked up from a handful of attributes:

• Card type and tier. Debit versus credit, standard versus premium versus commercial. The single biggest factor.
• Merchant category. Supermarkets, fuel, charities and utilities often get preferential rates; categories with high dispute rates pay more.
• Channel and security. Card-present with chip is cheaper than card-not-present; transactions with strong authentication can qualify for better treatment than raw keyed entry.
• Geography. Domestic, intra-regional and inter-regional transactions carry different schedules.

Two practical consequences. First, a merchant can influence its rate mix at the margin, by qualifying transactions properly, submitting the data that better rates require, and using the security features the grid rewards. Sloppy integrations literally pay a higher price per transaction. Second, when a payment provider quotes one blended rate, it is averaging across this grid, and the average conceals which of your transactions are subsidising which.

Once interchange is visible, several industry structures decode themselves.

Why do fintechs love issuing cards? Because interchange is revenue that arrives with every swipe of someone else's money. A neobank whose customers spend through its debit card earns a slice of every purchase, which in lightly-capped markets can carry a business model on its own, and in capped markets pushes issuers toward commercial cards, where caps do not bind.

Why do merchants build their own payment methods and push bank-transfer checkouts? To route around the stack entirely. Every account-to-account payment is a transaction on which no interchange is paid.

Why did buy-now-pay-later providers charge merchants 3 to 6 percent and find them willing? Because the comparison point was not zero, it was card fees plus the conversion uplift, and the providers took on the credit risk that justified an issuer-sized cut.

The general lesson: in payments, the fee table is the org chart. Knowing who receives each basis point tells you who will build, block or lobby for what, more reliably than any press release.