Fintech Cases 2026: A Take Rate Set by the Regulator

A revenue line written in a rulebook

A payments business earns a take rate on volume it does not own. Revenue equals transaction count times revenue per transaction, and the second term is where the sector stops resembling anything else.

For debit cards issued in the United States by a bank above a size threshold, the revenue per transaction is capped. A covered issuer "may not receive, for any electronic debit transaction, an interchange fee that exceeds $0.21 plus 0.05 percent multiplied by the value of the transaction, plus a $0.01 fraud-prevention adjustment, if eligible", checked 30 September 2026 at federalreserve.gov.

Read the shape of that formula, because it dictates the whole economics. Twenty one cents is flat. Five basis points is proportional. One cent is conditional. At an average ticket the ceiling works out at 24.3 cents, and the flat 21 cents is 86 percent of it, so most of the revenue stands still when the ticket moves.

The cost side has three lines, and all three are published. Processing cost per transaction. The issuer's share of fraud losses. Network fees. Subtract them from the capped revenue and you have a per transaction margin built entirely from public numbers.

Start by checking the public numbers against each other. In 2023 the networks "processed 100.7 billion debit and general-use prepaid card transactions valued at $4.7 trillion", and interchange across all of them "totaled $34.12 billion", checked 30 September 2026 at federalreserve.gov. So the average ticket is 4,700 divided by 100.7, which is $46.67. And average interchange is 34.12 divided by 100.7, which is 33.9 cents, which reproduces the 34 cents the Federal Reserve publishes itself. The two series agree, and that test takes ten seconds.

Brightpath, a fintech case priced to the tenth of a cent

Brightpath runs a debit program through a sponsor bank above the size threshold, so every transaction is covered. It forecasts 240 million transactions next year at the market average ticket. Fixed cost is $22 million. The question is whether to launch.

Build the revenue per transaction first. 0.05% on $46.67 is $0.0233. Add the flat 21 cents and the one cent adjustment and the cap is $0.2433. Covered issuers actually received 23 cents on average in 2024, and in 2023 the average was 24 cents on single-message networks and 22 cents on dual-message networks. Use the observed 23 cents rather than the ceiling, and say why out loud.

Now the three cost lines, all for 2023.

1. Processing. The average cost of authorization, clearance and settlement was 4.1 cents overall and 3.6 cents for high volume issuers. Brightpath is high volume, so use $0.036.
2. Fraud. Losses to all parties were "17.6 basis points, or $17.63 per $10,000 in transaction value", and issuers "absorbed 28.3 percent of losses". So the issuer's share is 0.283 times 17.6, which is 4.98 basis points. On a $46.67 ticket that is $0.0232.
3. Network fees. The average network fee was 12.9 cents a transaction, and "acquirers and merchants paid 64.9 percent of network fees; issuers paid the rest". The issuer's 35.1 percent share is 0.351 times 0.129, which is $0.0453.

Subtract in order. From 23 cents take 3.6 cents, leaving $0.194. Take 2.32 cents of fraud, leaving $0.1708. Take 4.53 cents of network fees, leaving $0.1255.

At 240 million transactions, 240 million at $0.1255 is $30.12 million of contribution. Against $22 million of fixed cost the program earns $8.12 million. Breakeven volume is 22 divided by 0.1255, which is 175.3 million transactions, so the plan clears its threshold with 27 percent of headroom.

Why a bigger basket adds almost nothing

Raising average transaction value is an obvious lever to reach for. On a covered debit program it is worth two multiplications before anyone builds a plan around it.

Take the ticket from $46.67 to $70.00, a rise of $23.33. The ad valorem slice of the cap is five basis points, so revenue rises by 0.0005 times 23.33, which is $0.0117. The issuer's fraud slice is 4.98 basis points, so fraud cost rises by 0.000498 times 23.33, which is $0.0116.

The two moves cancel. Net gain per transaction is about five thousandths of a cent, on a 50 percent rise in basket size.

That result is not a coincidence, and it is worth stating precisely. The proportional part of the regulated revenue happens to sit almost exactly on top of the proportional part of the issuer's fraud exposure: 5.00 basis points against 4.98. So on covered debit, growth has to come from transaction count. A merchant mix shift toward larger baskets moves the top line and leaves the contribution where it was.

The same arithmetic runs the other way for a credit program, where interchange is mostly ad valorem and the flat component is small. A candidate who carries the debit conclusion into a credit case will be wrong by the full width of the basket effect, and the interviewer will be watching for exactly that transfer.

The ten billion dollar line in the balance sheet

The exemption is the largest single number in this sector and it has nothing to do with the product. "A small debit card issuer is an issuer that, together with its affiliates, has assets of less than $10 billion." Below that line, the cap does not apply.

The difference shows up in the published averages. Exempt transactions carried average interchange of 51 cents in 2024 against 23 cents for covered ones. In 2023 exempt dual-message transactions averaged 62 cents, down from 64 cents in 2021.

Price the threshold for Brightpath. On 240 million transactions the gap between the 51 cent exempt average and the 23 cent covered average is 240 million at $0.28, which is $67.2 million of interchange. Both figures are for 2024 and come from the same table, which is the only way the subtraction means anything.

Now say what that shows. Brightpath's entire contribution is $30.12 million. Its sponsor bank crossing the $10 billion asset line would cost the program $67.2 million of interchange, which is more than twice everything the program earns. The sponsor choice is therefore a larger lever than pricing, merchant mix, fraud tooling and basket size combined, and it belongs in the structure at the top rather than in a risks section at the end.

Two conditions limit that comparison, and both should be said. The exempt average of 51 cents is taken across a different mix of issuers and merchants than Brightpath's own, so it is a market benchmark rather than a forecast. And the biennial collection that produced these figures is required of covered issuers by the Electronic Fund Transfer Act, checked 30 September 2026 at federalreserve.gov, so the 2023 series is the most recent full year on that page.

Where a payments answer goes wrong

Our own grader puts four weights on a piece of case arithmetic. They are our weights, taken from the rubric code in this repository, and no card network or consulting firm has endorsed them. What sits behind each one is set out on how an answer is assessed.

Drilling the per transaction chain in a fintech case

Rework Brightpath with one line moved at a time, in cents, and name the unit each time.

Fraud first, because it is the line a fintech can actually change. Halve the issuer's fraud share from 4.98 to 2.49 basis points. Fraud cost per transaction becomes 0.000249 times 46.67, which is $0.0116, so contribution rises to $0.1371. On 240 million transactions that is $32.90 million, worth $2.78 million a year. Breakeven volume falls to 22 divided by 0.1371, which is 160.5 million transactions.

Network fees second, because they are negotiated rather than fixed. Take the issuer's share of the 12.9 cent network fee from 35.1 percent to 30.0 percent. The cost becomes 0.30 times 0.129, which is $0.0387, so contribution rises by $0.0066 to $0.1321. That is $1.58 million a year on 240 million transactions, which tells you roughly what a network renegotiation is worth before you spend a quarter pursuing one.

Then check the assumption underneath all of it. Every figure here is a market average for calendar year 2023, and a single program's mix of card-present and card-not-present volume shifts both the ticket and the fraud rate. Card-not-present transactions averaged $63.96 in 2023 against $36.99 for card-present, so a program weighted toward one or the other is working from a different base entirely.

Our breakeven calculator does the fixed cost division above, and the math drill times chains like this one and returns Answer, Working, Units and Interpretation separately. Both are free and need no account.

For the two rate version of financial economics, see banking cases. For fee revenue on an asset stock rather than on volume, see financial services cases. For a case where the volume itself has to be sized first, see the market entry framework.