The Durbin Amendment caps debit interchange fees at roughly 22-24 cents per transaction for banks with over $10 billion in assets — but only for “four-party” systems like Visa and Mastercard. “Three-party” systems, where the same company is both card issuer and network, are statutorily exempt — a carve-out written for American Express and Discover, never for a bank that just bought one.[1] Capital One's $50.6 billion acquisition of Discover, completed May 18, 2025, did exactly that: migrate debit cards onto Discover's uncapped network (roughly 1.10%+16 cents per transaction, versus the capped 22-24 cents) instead of Visa's or Mastercard's.[2] Four consecutive earnings calls tracked the migration from pilot to completion — Capital One's own CEO stated on April 21, 2026 that “the debit conversion is completed.”[2] And yet: no isolated dollar figure for the interchange gain has been disclosed. No merchant or trade group has gone on record with real, measured cost data — the widely-repeated “merchants charged up to 3x more” claim traces to a payments consultancy's hypothetical modeling exercise, not a real bill.[3] No regulator has followed up since the deal closed. A workaround executed on schedule, confirmed by the company's own words, and its actual consequences remain entirely undisclosed a year later.
The Durbin Amendment, part of the 2010 Dodd-Frank Act, caps debit interchange fees for large banks at a base of 21 cents plus 0.05% plus a 1-cent fraud adjustment — roughly 22 to 24 cents on a typical transaction. But that cap only applies to “four-party” payment systems, where issuer, network, acquirer, and merchant are separate parties. “Three-party” systems — where one company is both the card issuer and the network — are exempt by statute. That exemption was written for American Express and Discover. It was never written with the expectation that a capped bank would simply buy one.[1]
Capital One did exactly that. Its $50.6 billion acquisition of Discover Financial Services closed May 18, 2025, and the strategic logic was explicit from the start: migrate Capital One's debit cards onto Discover's network, where interchange runs uncapped — roughly 1.10% plus 16 cents on card-present transactions, versus the capped 22-24 cents. Four consecutive quarterly earnings calls tracked the migration in increasingly definitive language: a June 2025 pilot, “nearly complete” by January 2026, and finally, on the April 21, 2026 call, CEO Richard Fairbank stated plainly that “the debit conversion is completed.” Independent SEC filing data and third-party transaction-volume tracking corroborate the timeline.[2]
Here is what hasn't happened: Capital One has never disclosed an isolated dollar figure for what this specific migration is worth. Its consolidated interchange-fee revenue did rise sharply — but the company doesn't break out how much reflects the Discover-network shift versus simply having a much larger combined card base post-merger. Its own 10-K risk-factor language hedges further: if Discover-routed debit were ever subjected to the Durbin cap, the company could lose “a significant majority” of the network revenue synergies it has been anticipating — an admission, in its own filing, that the gain isn't secured.[3]
Nor has anyone quantified the cost on the other side. The specific claim that merchants are now being charged “up to 3x” more traces to a hypothetical modeling exercise by a payments consultancy — a modeled $50 transaction, a hypothetical $1 billion-volume merchant — not a survey, a named retailer, or real billing data. No major merchant group has gone on record with measured, post-migration cost figures. Senator Elizabeth Warren wrote to the Federal Reserve and DOJ flagging the exemption and a ~$1.2 billion revenue opportunity — but those letters are dated May 1 and 13, 2025, before the deal even closed, and no regulator has followed up since. And on July 6, 2026, JPMorgan, Bank of America, Wells Fargo, and PNC were reported in early talks to buy Fiserv's STAR and Accel debit networks — the same playbook, one year later, still too fresh to have drawn a single on-record reaction from any party.[4]
How a fee-cap workaround moved from strategy to confirmed completion, without a single disclosed dollar figure.
Dodd-Frank's Durbin Amendment caps four-party debit interchange fees for large banks, but exempts three-party systems — a carve-out for American Express and Discover, not anticipating a capped bank acquiring one.[1]
The RuleThe $50.6 billion acquisition closes. The strategic logic, stated from the start: migrate debit volume onto Discover's uncapped network.[2]
The AcquisitionFour consecutive earnings calls chart the conversion from pilot to “nearly complete” to, finally, Fairbank's April 2026 statement: “the debit conversion is completed.”[2]
ConfirmedNo isolated interchange-revenue figure has been disclosed. The widely-cited “3x merchant cost” claim turns out to be a consultancy's hypothetical model, not a real bill.[3]
The GapJPMorgan, Bank of America, Wells Fargo, and PNC are reported in early talks to acquire Fiserv's STAR and Accel debit networks — the same three-party workaround strategy, one year later, still unconfirmed.[4]
RecurringThe debit conversion is completed. — Richard Fairbank, CEO, Capital One, Q1 2026 earnings call, April 21, 2026
| Dimension | Evidence |
|---|---|
| Revenue (D2) Origin · 86 | The lever is a fee-capture mechanism: restructuring which regulatory category a bank's debit transactions fall under, via ownership of a three-party network.[1][2] D2 is the origin because nothing about the underlying purchase or cardholder behavior changed — what changed is which interchange rate applies to the identical transaction.The Fee-Capture Mechanism |
| Operational (D6) L1 · 82 | The operational fact is unusually well-confirmed: four consecutive earnings calls, cross-checked SEC filings, and independent transaction-volume tracking all agree the migration is complete as of April 2026.[2] D6 amplifies from D2 because it's the concrete proof the mechanism was actually executed, not merely proposed.Confirmed Complete |
| Quality (D5) L1 · 80 | No isolated revenue figure has been disclosed for the migration itself, and the loudest counter-claim about merchant harm turns out to be a hypothetical model, not measured data.[3] D5 amplifies alongside D6: the mechanism is confirmed, but the honesty of what it's actually worth — to the bank or to merchants — remains an open gap on both sides.The Honesty Gap |
| Customer (D1) L2 · 60 | Merchants and, downstream, consumers are the parties who would bear any real cost increase — but no trade group or named retailer has produced real measured data confirming or denying that cost has changed.[3] D1 sits here as the dimension where the absence of evidence is itself the finding. |
| Regulatory (D4) L2 · 56 | Senator Warren's warnings to the Fed and DOJ preceded the deal's close; no confirmed regulatory follow-up has occurred since completion.[4] D4 sits at a middling score because the regulatory apparatus engaged early but has gone quiet exactly when the migration finished. |
| Employee (D3) 38 | Deliberately the thinnest dimension. This is a fee-structure and disclosure cascade; there is no comparable workforce-level finding in the research to build a dimension around. |
The cascade originates in D2 — Revenue — because the lever is a fee-capture mechanism: a bank restructuring which regulatory category its transactions fall under, not a change in the underlying transactions themselves.[1][2] From D2 it amplifies into D6 (the operational reality — a real, confirmed, four-earnings-call migration) and D5 (the honesty gap — no isolated revenue figure disclosed, and the loudest merchant-harm claim is modeled, not measured).[2][3] It then reaches D1 (merchants and consumers, where no real cost data exists either way) and D4 (regulatory response — pre-close letters, no post-close follow-up).[4] D3 is thin — this is a fee-structure cascade, not a workforce one. Cross-references: [UC-267] is the parallel legal threat to the entire capped regime, running independently of this corporate workaround; [UC-268] is the necessary counter-cascade — why this drew none of the backlash a similar move drew in 2011; [UC-269] holds open whether either track resolves the money question this case cannot yet answer.
-- UC-266: The Fee Nobody Can Find: 6D Diagnostic Cascade
-- Capital One-Discover debit migration confirmed complete, financial impact undisclosed (cluster: UC-267/268/269)
FORAGE fee_nobody_can_find
WHERE three_party_exemption_used = true
AND migration_confirmed_complete = true
AND isolated_revenue_figure_undisclosed = true
ACROSS D2, D6, D5, D1, D4, D3
DEPTH 3
SURFACE fee_nobody_can_find
DIVE INTO workaround_without_a_number
WHEN mechanism_executes_on_schedule = true
AND consequences_remain_undisclosed = true
TRACE invisible_impact_cascade
EMIT fee_arbitrage_signal
DRIFT fee_nobody_can_find
METHODOLOGY 86
PERFORMANCE 44
FETCH fee_nobody_can_find
THRESHOLD 1000
ON MONITOR CHIRP high 'Capital One's Discover acquisition let it route debit cards through the three-party exemption to the Durbin Amendment's interchange cap. Migration confirmed complete April 2026. No isolated revenue figure disclosed; the 'merchants charged 3x more' claim is a consultancy's hypothetical model, not real data; no regulator has followed up since close. JPMorgan/BofA/Wells/PNC reportedly exploring the same playbook via Fiserv's STAR/Accel networks, July 2026'
SURFACE analysis AS jsonRuntime: @stratiqx/cal-runtime · Spec: cal.semanticintent.dev · DOI: 10.5281/zenodo.18905193
Four consecutive earnings calls, cross-verified by independent filings and transaction data, confirm the migration happened exactly as planned. What's genuinely missing — from the bank, from merchants, from regulators — is any disclosed number for what it actually changed.[2][3]
The “3x higher merchant costs” figure repeated across coverage traces to a consultancy's modeled $50 transaction, not a survey or a named retailer's real bill. It's a clean example of a plausible-sounding number outrunning its own evidence.[3]
Capital One's 10-K admits that if regulators ever extended the cap to Discover-routed debit, it would lose “a significant majority” of the anticipated synergy. Even the beneficiary of this workaround treats it as legally unresolved, not secured.[3]
Senator Warren's warnings landed in May 2025, before completion. Once the migration actually finished, in April 2026, the silence has been total — no regulator, no merchant group, no follow-up letter.[4]
Four sources: the statutory mechanism itself, Capital One's own earnings-call and SEC-filing record of the migration, the debunked hypothetical-modeling origin of the merchant-harm claim, and the fresh, still-unconfirmed Fiserv reporting showing the same playbook recurring.
Confirmed mechanism, confirmed completion, undisclosed consequence. That gap is the actual story.