Capital One has collapsed US debit to three networks

The Capital One acquisition of Discover closed on 18 May 2025. The Federal Reserve approved the deal under Order 2025-10 on 18 April, carrying a simultaneous $100 million consent-order fine against Discover Bank for interchange overcharges from 2007 to 2023; the OCC issued conditional approval the same day (Federal Reserve press release on the Order; Capital One close announcement). The combined entity holds roughly 22% of US credit-card balances. The merger has been written about almost entirely as an issuer-side story.

That framing has undertold the binding consequence. The merger ended the four-network US debit market. Capital One has migrated roughly $175 billion of annual debit volume onto Discover’s three-party rail, taking the largest non-network-owned debit issuer off Pulse-as-shared-utility and absorbing it into the network owner. Mastercard has publicly conceded adverse 2026 and 2027 impact; Visa has called the migration a small factor. The consequence sits on the debit side, while the antitrust review concentrated on credit-card overlaps rather than network competition. The clearest read of the resulting math is not from the large-issuer balance sheet but from the small-issuer one.

The numbers are unambiguous. The Federal Reserve published its 2023 biennial debit interchange survey on 19 December 2025. Total US debit interchange across covered and exempt issuers reached $34.12 billion. Covered issuers (with $10 billion or more in assets, subject to the Durbin cap) earned $14.02 billion at $0.23 per transaction. Exempt issuers (under $10 billion in assets, not subject to the cap) earned $20.10 billion at $0.52 per transaction (Federal Reserve press release on the 2023 biennial; Fed report page). Exempt small issuers earn the larger absolute share of US debit interchange at more than twice the per-transaction rate of covered issuers. The exempt rate is a deliberate carve-out in Reg II, not a loophole; closed-loop networks (Discover, American Express) and small issuers were intentionally exempted because the cap regulates interchange paid between two unaffiliated parties, and closed loops have no interchange in that sense, while small issuers were exempted on the policy view that the cap would disproportionately harm institutions with smaller transaction volumes.

The asymmetry is what makes the small-issuer seat the right vantage. A 10 to 15% compression of debit interchange revenue at JPMorgan is absorbed across a $4 trillion balance sheet. At a $200 million CDFI credit union, debit interchange is a primary revenue line. The covered-issuer balance sheet has fee, lending, and trading lines that diversify away the policy variable, while the exempt-issuer balance sheet does not. The seat with the highest income elasticity to debit interchange is therefore the seat from which the merger’s redistribution math reads cleanest.

The merger’s network-economic harm was identified by the agency closest to the analytic question and then politically dismissed. The Department of Justice Antitrust Division’s January 2025 staff memo, drafted under the Biden-era Antitrust chief, found the deal anticompetitive on debit, noting that Capital One would use the merger to avoid interchange fees on its own debit cards (Capitol Forum on the staff memo). In April 2025, with Antitrust chief Gail Slater in place, the Division reversed and issued a confidential memo to the Federal Reserve and OCC stating it had insufficient evidence to block (Finextra on the April 2025 reversal). The Federal Reserve approved on the credit-card analysis, the OCC issued conditional approval, the deal closed. The reversal did not refute the staff analysis; it declined to act on it. The staff diagnosis identified a real network-economic effect, in which Capital One would bypass the Durbin cap on its own debit volume by routing through the Discover three-party rail, and quantified the uplift at roughly $1.2 billion. That diagnosis now sits unresolved between the agencies and the merged firm.

Capital One has been precise about the strategy on its own earnings calls. Pre-close, founder and CEO Richard Fairbank described Pulse as the rare asset in the deal and framed the migration target: 25 million debit cards and roughly $175 billion of annual purchase volume by 2027 (Digital Transactions on Pulse strategy). On the Q4 2025 call, Fairbank reported the migration substantively complete: “We’ve been migrating our debit card holders to the Discover network since August” (Capital One Q4 2025 transcript). Credit migration is forward-looking; origination on Discover begins mid-2026, with portfolio migration following. Management has guided roughly $2.5 billion to $2.7 billion in total synergies by 2027, of which $1.2 billion is network synergy tied to debit migration and selected credit migration onto Discover and Pulse rails. The transfer is from the merchants paying Discover-network rates rather than Visa or Mastercard interchange-capped rates on the same volume.

The schemes’ own earnings disclosures calibrate the magnitude. Mastercard CFO Sachin Mehra on the Q3 2025 call: “The debit migration is underway. No surprise there” (Mastercard Q3 2025 transcript). On Q4 2025 he was more specific: “The growth of our debit portfolio was impacted by the Capital One debit migration” (Mastercard Q4 2025 transcript). The expected adverse impact runs into 2026 and 2027, partially mitigated in 2026 by contractual offsets and unmitigated in 2027 once those obligations expire. Mastercard renegotiated the credit-card agreement with Capital One pre-close to preserve the credit relationship (Payments Dive on the renegotiation). Visa has been muted by comparison. CFO Christopher Suh on the Q1 fiscal-year 2026 call attributed a slight US payments-volume step-down to debit, including “the loss of some interlinked volumes to the Capital One debit migration,” framed as one of “a number of other small factors” (Visa Q1 FY2026 transcript). The two framings differ in tone; both schemes have publicly conceded a real debit-volume migration. The four-network US debit market has gone to three.

The legal context that determines how durable the re-pricing turns out to be is itself unstable. Regulation II has bifurcated. The District of North Dakota in Corner Post v. Federal Reserve, decided 6 August 2025, vacated Reg II’s interchange-fee cap and then stayed its own vacatur pending Eighth Circuit appeal. The court held that the Federal Reserve had impermissibly included fixed costs, network fees, and fraud losses in the cap calculation when only incremental costs were authorised; that the uniform cap violated the statutory requirement to assess costs with respect to the transaction; and that the rulemaking displaced courts as arbiters of statutory meaning (Cooley advisory on Corner Post; ABA on Fed reply brief). The Federal Reserve filed its reply brief on 23 March 2026; oral argument has not been scheduled. The Corner Post challenge attacks the fee cap, not the routing rule. The 2022 Reg II amendment extending the two-unaffiliated-network requirement to card-not-present transactions remains operative. Meanwhile, the Federal Reserve’s 2023 NPRM proposing a cap reduction from $0.21 to $0.144 base, with adjusted ad valorem and fraud-prevention legs, has not advanced. Then-Governor Michelle Bowman dissented from the NPRM in October 2023, arguing the cap risked putting “smaller issuers at a significant competitive disadvantage” (Bowman October 2023 dissent). Her elevation to Vice Chair for Supervision in June 2025 is the proximate reason the NPRM is unlikely to advance in current form.

Two further pieces sit on the policy track. The Credit Card Competition Act, reintroduced in January 2026, would extend the two-unaffiliated-network rule to credit. ICBA, in a July 2025 letter to Vice Chair Bowman, argued that the proposed cap would leave a third of covered issuers below cost on debit (ICBA July 2025 letter). Doug Kantor, NACS general counsel and a Merchants Payments Coalition executive committee member, called the December 2025 Fed biennial “the consequences of the Federal Reserve failing to keep up” (Merchants Payments Coalition statement). The legal frame is unstable; the merchant and small-issuer voices are consistent.

The contested ground is sharper than the political middle. The ICLE white paper on the merger is the cleanest pro-merger position to engage. Eric Morris, Eric Fruits, Lazarus Sperry, and Todd Zywicki argued that by moving customers onto Discover’s three-party network, Capital One’s customers will “avoid the distortions imposed by the Durbin Amendment’s price controls” (ICLE white paper). Their framing is that the merger strengthens competition by giving Discover the scale to challenge Visa and Mastercard, and that the resulting interchange uplift funds debit-rewards programmes that benefit lower-income consumers. Patrick Woodall, writing for Americans for Financial Reform in July 2024, framed the same fact pattern from the opposite side: the deal lets Capital One shift its debit cards onto a network charging “interchange fees that are double” the covered-issuer rate, projected to generate $1.2 billion in higher revenues (Americans for Financial Reform top-10 reasons). Both positions cite the same Durbin three-party exemption; they differ on whose welfare matters at the margin.

The CDFI-board seat reframes the contest. The piece most economists ignore is that the Durbin exemption-bypass is recovered from the same merchant pool that pays the small-issuer exempt rate. The merger is not a transfer between Capital One and Visa or Mastercard. It is a transfer from merchants to Capital One that does not move any of the exempt-issuer revenue. Exempt-rate small issuers continue to receive their $0.52 per transaction and covered-rate large issuers continue at $0.23, while Capital One leaves the covered cohort and joins the closed-loop network at rates closer to the exempt; the merchant pool absorbs the difference.

The second-order consequence the CDFI-board seat sees first is the Pulse community-bank competition question. Pulse has roughly 4,500 community-bank and credit-union issuer relationships, and those issuers now have a direct issuer-competitor as their network owner (Digital Transactions on Pulse positioning; The Financial Brand summary). The economic analogue is owning a marketplace while running the largest seller on it; whether community banks remain on Pulse becomes a function of their willingness to share routing data with a competitor rather than of Pulse’s pricing. Andrew Dresner’s pre-close analysis surfaced the same point through a different lens. Pulse’s competitive position in PIN debit is disadvantaged without federation with Chase or Bank of America, and “exempt interchange status [is] the most compelling network ownership benefit” (Dresner, Payments in Full, March 2025). The strategic move that maximises Capital One’s exempt-status capture is the same move that makes the underlying Pulse asset less attractive to its community-bank issuers, because the network owner has become their largest competitor.

The Federal Reserve Bank of Philadelphia’s Working Paper 25-18 (Hunt, Serfes, Zhang, June 2025) models interchange in two-sided card networks and finds that capping interchange benefits all consumers by lowering the endogenous credit-card tax, even if rewards decrease (Philadelphia Fed WP 25-18). The paper does not address the merger directly, but its model implies that the cap was holding down a tax on consumers paid through merchants; vertical integration that bypasses the cap reverses the welfare direction the model identifies. ICLE has rebutted the model on empirical grounds (ICLE rebuttal). Both sides use the same economist toolkit; the contested ground sits between equilibrium pricing under bypass and welfare under cap retention.

A reader can object that the merger is mostly an issuer-side story and the network effect is over-attributed; the DOJ cleared on credit-card analysis, Visa calls the migration a small factor, and ICLE argues the merger strengthens competition by giving Discover the scale needed to operate as a credible third rail. The calibration is fair, the conclusion is not. The DOJ staff memo identified the network harm before the political reversal. Small factor is ex-ante volume guidance and tells us nothing about the rate-card consequences. The stronger-third-competitor frame concedes the four-to-three shift on debit; it disputes only whether the welfare effect is positive or negative. The migration itself is real even where its competitive magnitude is contested.

To the further objection that the framing should be three-and-a-half rather than three, because American Express operates a closed-loop, Durbin-exempt rail: conceded on credit. Amex is similar in shape to Discover-Pulse-Capital-One, and US credit volume across the four open and closed networks remains roughly three-and-a-half. The three-network framing applies precisely to debit. Amex’s US debit volume is rounding error against the roughly $4.7 trillion total US debit value, so the article’s claim of three debit networks rather than four is precise as stated.

The remaining question is whether Capital One has actually moved the volume yet. Migration is mostly debit and mostly already done by Q1 2026; credit migration is slower, with origination on Discover beginning mid-2026 and portfolio migration following. The credit story is forward-looking and the debit story is already being scored. Mastercard’s Q3 and Q4 2025 calls publicly disclose adverse 2026 and 2027 impact, Visa’s Q1 FY2026 call discloses lost interlinked volumes, and Capital One’s Q4 2025 call confirms migration substantively complete. The debit reset is a 2026 event, not a 2027 forecast.

The closing question is who benefits and who absorbs the cost. Private-equity diligence on US debit-issuance assets through 2027 must press not whether the asset’s interchange revenue holds at current levels but whether its competitive position holds in a market where the largest non-network-owned debit issuer has joined the closed loop. Pulse-issuance contraction risk is real for any community-bank or credit-union asset, and acquirer-side assets that route through Pulse face the network-owner-as-competitor question on the routing side. CCCA passage would partially restore competitive routing on credit but does not address the debit re-pricing already in progress. Supervisory attention at the Federal Reserve and OCC carries a different question: whether the staff-memo analysis politically dismissed in April 2025 is reflected in the post-merger supervisory frame. The Fed’s biennial debit survey is the audit, and the next survey, covering 2025 data, will cover the year of the merger close. The most direct exposure sits at CDFI and community-bank boards, where the institution issues on a network now owned by its largest competitor and retains less routing optionality than at any time since 2011. The decisions that follow are concrete: whether to reprice the issuance partner, renegotiate Pulse terms, or accept that the institution’s primary revenue line is being recovered by an entity competing with it on issuance.