Reg II 2026 Debit: Fed Cost Trail, Square-Adyen, Caveats

The catch: the Fed does not send anyone a check. Interchange moves from issuer to acquirer; what happens next is set entirely by the merchant's processing agreement. Flat-rate platforms — Square and Stripe's standard pricing lead the list — quote a single blended percentage with no reason to move when the capped component beneath it falls. Itemized, interchange-plus models in the Adyen mold reprice with the cap. Same regulation, opposite outcomes, turning on one clause most owners have never read: whether interchange is itemized.

Hence the real 2026 enforcement story: not Washington versus banks, but merchants versus their own contracts. The networks carry the mandate — Visa's core rules bind issuers and acquirers alike — yet delivery to a P&L is contractual. Pull 12 months of statements and look for interchange as a line item; without it, the 10 bps is spoken for, and renewal clocks on 12- or 24-month terms decide when that changes.

Content for Inside the Cap is being prepared.

Inside the Cap

Decision 1 — timing a card application. Priya wants the Bank of America Air France/KLM Flying Blue Mastercard ahead of a fall trip to Paris, but she closed her previous Flying Blue Mastercard in November 2024. Bank of America's 24-month rule requires waiting 24 months after either opening or cancelling the same card before becoming eligible again, putting her earliest realistic window at November 2026. Because she holds a BofA deposit account, the alternative standard applies — no more than 7 new card accounts in the past 12 months instead of the stricter default — and with only two new accounts opened, she clears it easily. Forum reports note that enforcement seems to vary, so she schedules the application for early November rather than counting on flexibility.

The Receipt Trail

The statute behind them is Dodd-Frank, which requires debit interchange to be "reasonable and proportional" to the issuer's processing costs. The Fed's surveys measuring roughly 5¢ of per-transaction authorization, clearing, and settlement cost against 24.5¢ charged are the documented disproportion the revised formula is built to close. This matters commercially: when an acquirer insists the savings "aren't material," your reply is a statutory requirement backed by the regulator's own cost data.

History says the repricing will be abrupt, not gradual. According to the Federal Reserve's first biennial review, covered-issuer average interchange collapsed from roughly 44¢ to about 24¢ when the original cap took effect. Formula changes translate into immediate, statement-visible repricing. Expect the same shape in 2026: the new fee level appears on statements right after the effective date, but only on statements that itemize interchange in the first place.

Why did a proposal shelved in 2023 become 2026 enforcement? Corner Post v. Federal Reserve, decided unanimously by the Supreme Court on July 1, 2024. The Court held that the six-year statute of limitations runs from injury, not from the original rule's enactment — reviving challenges to the original cap's level and converting the dormant proposal into live regulatory pressure. The timing was not administrative housekeeping; it was litigation-driven.

Now the scope check, and the trap. According to the Fed's 2022 interchange report, roughly two-thirds of U.S. debit transaction volume runs through covered issuers — the reset directly reprices that majority while the exempt third continues pricing without limit. So kill this myth now: a cut of this size in a fee line is not a cut of this size in your bill. The cap binds only covered issuers and only the interchange line, leaving acquirer markup, network switch fees, and the fraud add-back entirely outside it. A flat-rate merchant will see zero change on the compliance date.

The concrete move: pull twelve months of statements, identify your covered-issuer debit share, and put one question to your acquirer in writing — which line on this statement changes on the 2026 effective date, and by how much per transaction? If the answer is "your rate," you are on blended pricing, and the receipts above become your argument for the interchange-plus structure with least-cost dual routing that lets these figures land as a lower line instead of absorbed margin.

Run the same covered-debit ticket through Square and through Adyen after the compliance date and only one statement changes. That asymmetry is the whole scoreboard. Reg II caps the issuer-side interchange input, so a cost-plus contract reprices mechanically while every blended structure keeps the difference. According to the Medium analysis of interchange-plus pricing, cost-plus runs both more transparent and generally cheaper than tiered billing — and the revised cap widens that gap, because the wholesale base under your markup is the one line the Fed controls. It also kills the myth that the cut shrinks every processing bill: on flat-rate pricing, the compliance date moves nothing.

Pricing is half the capture problem; routing is the other half. Regulation II lets a covered transaction traverse any enabled network, and least-cost rules steer volume to Star/Accel-class rails that typically undercut the default Visa/Mastercard path — so two quotes with identical markups realize different basis points depending on who controls the switch. On legacy terminals the routing decision usually is not yours at all; on modern gateways it is a settings page.

The trap row is flat-rate. A one-line quote — 2.89% + 10¢, the standard Square-style headline — wins on readability and loses on mechanics: issuer interchange falls, the blended rate does not move, and the spread widens by construction. The statements tell the story:

Acquirer Scoreboard

When two interchange-plus quotes compete, four checks separate pass-through from theater: compute the effective rate on your own last-90-days transaction file, never the advertised rate; demand per-item switch and network fees disclosed separately, not folded into the markup; verify dual routing on card-not-present credentials with a live test transaction, not a sales deck; and require contractual language promising dollar-for-dollar repricing whenever Reg II parameters change, fraud add-back included — the clause that makes the next revision automatic too.

The table forces one of two moves. On flat-rate or tiered pricing, the 2026 reset is a renegotiation event with quantifiable stakes — roughly 10 bps of covered-debit volume, reachable only through itemized cost-plus with least-cost dual routing. Already interchange-plus, the job is verification, not shopping: pull the first statement after the compliance date and confirm the covered-debit interchange line repriced downward. If it did not, the contract never carried true pass-through — fix that before renewal, not after.

Everything upstream of this section is a forecast wearing a measurement's clothes. The cap formula itself is settled law; what lands on any individual merchant's statement after enforcement is not, because the savings band came out of the Federal Reserve Board's regulatory impact analysis as a modeled blend across issuer fraud tiers — not read off a single post-compliance statement, of which none exist yet. Hold that thought against the seductive myth arithmetic: the cap falls sharply, so the processing bill must fall sharply too. That fails twice. The cap reaches only covered issuers — roughly two-thirds of debit volume — and only the interchange line, leaving acquirer markup, network switch fees, and the fraud add-back priced entirely outside it. A flat-rate merchant sees zero movement on the compliance date.

Three evidentiary gaps deserve honesty. First, the underlying cost data are issuer-reported: the Fed's Survey of Issuer Costs of Debit Card Processing aggregates self-reported figures from a sample of institutions, and averages built that way conceal wide per-issuer dispersion — two covered banks with identical ticket profiles can sit far apart once fraud losses, tokenization adoption, and network portfolios differ. Second, the acquirer side is opaque by design: regional-network switch fees at Star, Accel, and Shazam are negotiated privately and appear on no public schedule, so "least-cost routing" is computed inside your processor's engine, verifiable only if your contract forces line-item disclosure. Third, attribution: network fees update on their own calendars, so even a perfectly itemized statement cannot isolate the cap's effect without a clean pre-compliance baseline.

Variance across cases is wider than any average suggests. Because the modeled benefit is denominated in basis points of ticket value, it scales with the ticket — fractions of a cent on a coffee, a dime-scale rounding error on a hundred-dollar cart, meaningful money only at ticket sizes most cafés never ring. Issuer mix matters just as much: the fraud add-back is assigned per issuer, so a merchant whose cards skew toward high-fraud-tier banks lands near the bottom of the band, while one drawing on low-fraud-tier issuers keeps more of the cut. Channel constrains too — PIN debit at the terminal routes freely, but wallet-tokenized e-commerce traffic can face network restrictions that quietly disable the cheaper leg.

What the Data Doesn't Tell You

So when does the rule break? Four conditions, each an edge case rather than a refutation. If your presentment is dominated by exempt issuers, the cap never binds and expected savings shrink toward whatever switch-fee differential routing alone recovers. If your processor charges a fixed per-item fee, network choice moves pennies while that fee dominates unit economics — least-cost routing cannot repair a bad per-item contract. If your traffic runs through device wallets, confirm empirically that the second network actually fires; token provisioning can override routing preferences. And if your "interchange-plus" statement nets network fees into one blended line, the decision rule is unverifiable by construction — paying a premium for genuinely itemized billing is justified only when the statement lets you audit the routing split ticket by ticket. The test that converts this model into your number costs nothing: pull twelve months of statements, compute your covered-versus-exempt share and ticket distribution, and make the processor demonstrate a dual-routed versus single-routed split on your real tickets before signing anything.

Content for What the Fed's Model Misses is being prepared.

A cap cut can only reach a statement line that exists. If your processor bills one blended rate, the revised Reg II formula never touches your P&L — the delta quietly widens processor margin instead. That is why the five rules below run in strict order: pricing structure first, routing second, measurement third, contract fourth, maintenance fifth. Execute them out of order and you will negotiate hard against a number you cannot see.

Rule 1 — Price before the date. Be on interchange-plus with itemized interchange rows before the Fed's 2026 compliance date. Test any quote with one question: show me Visa debit versus Star- or Accel-class debit, card-present versus card-not-present, as separate rows. A rep who cannot produce that breakdown is selling a blend, and a blend is where cap reductions go to die. This is also where the popular myth dies — the idea that a cap cut passes straight through to the bill one-for-one. The cap binds only covered issuers, roughly two-thirds of debit volume, and only the interchange line, so a flat-rate merchant should expect zero change on the compliance date.

What the Fed's Model Misses

Rule 2 — Route by default. Enable at least two unaffiliated debit networks on every card-not-present credential and set least-cost routing in the gateway so each authorization takes the cheapest qualified rail. Then audit monthly that the rules actually fired. In most gateway stacks, routing tables sit alongside terminal configuration, and a routine POS software update can silently restore defaults — transactions still clear, so nobody notices until the next statement review.

Harbor & Vine's Annual Recoverable Savings

Rule 3 — Model per BIN, not per average. Pull 90 days of statements, tag each issuing BIN covered versus exempt — asset-size exemptions mean some debit never caps, so tag rather than assume — and apply the 14.4¢ + 4 bps + fraud-tier formula transaction-by-transaction. Carry the result into negotiations as your baseline. If modeled recovery lands under roughly 5 bps of covered-debit volume, your problem is routing, not pricing: fix Rule 2 before shopping for a new acquirer.

Rule 4 — Contract the pass-through. Insert a rider committing the acquirer to reduce interchange billing dollar-for-dollar whenever Reg II parameters change, effective on the compliance date. Negotiate it at renewal, in writing, before the reset. The behavioral mechanics matter: once the lower interchange posts first, any later concession reads as processor generosity rather than regulatory obligation, and your renewal leverage evaporates with it.

Rule 5 — Re-underwrite quarterly. Track the fraud-adjustment tier distribution across your top issuing banks and the network fee lines on each statement — issuers recover fraud costs through tiered add-backs, so the surcharge you face varies with who issued the card. If an outsized share of covered volume migrates into the top 7¢ tier, or network fees creep upward post-reset, re-run the Rule 3 model and re-steer routing. Treat the ~10 bps reduction quantified earlier as a planning ceiling to be defended, not a permanent entitlement.

If you execute only one rule, execute Rule 1: pull tonight's statement and look for itemized interchange rows. Their absence means Rules 2 through 5 are optimizing a line that does not exist — and the entire cap reduction is currently someone else's revenue.

Five Rules to Lock In the Savings Before Your Processor

A cap cut can only reach a statement line that exists. If your processor bills one blended rate, the revised Reg II formula never touches your P&L — the delta quietly widens processor margin instead. That is why the five rules below run in strict order: pricing structure first, routing second, measurement third, contract fourth, maintenance fifth. Execute them out of order and you will negotiate hard against a number you cannot see.

Rule 1 — Price before the date. Be on interchange-plus with itemized interchange rows before the Fed's 2026 compliance date. Test any quote with one question: show me Visa debit versus Star- or Accel-class debit, card-present versus card-not-present, as separate rows. A rep who cannot produce that breakdown is selling a blend, and a blend is where cap reductions go to die. This is also where the popular myth dies — the idea that a cap cut passes straight through to the bill one-for-one. The cap binds only covered issuers, roughly two-thirds of debit volume, and only the interchange line, so a flat-rate merchant should expect zero change on the compliance date.

Rule 2 — Route by default. Enable at least two unaffiliated debit networks on every card-not-present credential and set least-cost routing in the gateway so each authorization takes the cheapest qualified rail. Then audit monthly that the rules actually fired. In most gateway stacks, routing tables sit alongside terminal configuration, and a routine POS software update can silently restore defaults — transactions still clear, so nobody notices until the next statement review.

Rule 3 — Model per BIN, not per average. Pull 90 days of statements, tag each issuing BIN covered versus exempt — asset-size exemptions mean some debit never caps, so tag rather than assume — and apply the 14.4¢ + 4 bps + fraud-tier formula transaction-by-transaction. Carry the result into negotiations as your baseline. If modeled recovery lands under roughly 5 bps of covered-debit volume, your problem is routing, not pricing: fix Rule 2 before shopping for a new acquirer.

Rule 4 — Contract the pass-through. Insert a rider committing the acquirer to reduce interchange billing dollar-for-dollar whenever Reg II parameters change, effective on the compliance date. Negotiate it at renewal, in writing, before the reset. The behavioral mechanics matter: once the lower interchange posts first, any later concession reads as processor generosity rather than regulatory obligation, and your renewal leverage evaporates with it.

Rule 5 — Re-underwrite quarterly. Track the fraud-adjustment tier distribution across your top issuing banks and the network fee lines on each statement — issuers recover fraud costs through tiered add-backs, so the surcharge you face varies with who issued the card. If an outsized share of covered volume migrates into the top 7¢ tier, or network fees creep upward post-reset, re-run the Rule 3 model and re-steer routing. Treat the ~10 bps reduction quantified earlier as a planning ceiling to be defended, not a permanent entitlement.