Mastercard (MA)

Total owner earnings compounded +17.2%/yr from FY21 to FY25. The market price requires +15.7%/yr for a decade. On height alone, the bar is cleared by the recent record. The tension is underneath: gross network assessments grew to $39,998M in FY25, but rebates and incentives — the payments back to issuers that buy the volume — grew faster, so the net yield (what Mastercard keeps of each gross fee dollar) fell from 51.0% in FY23 to 48.7% in FY25. If that slide is the toll for defending share, compounding continues but at a thinning margin on the core. If it accelerates, the delivered growth the bar leans on is being bought at a rising price. The October 29 print carries the first registered test of exactly that gap.

🐂 Bull case. A capital-light network compounding owner earnings in the mid-teens with a second engine — value-added services and solutions (VAS), $14,585M TTM, 41.6% of net revenue, growing near 20% — that monetises the same transactions again through security, authentication and data services. Cross-border volume (+12% local currency in Q2 2026) is the richest fee pool and is still compounding double digits. Buybacks shrink the count a demonstrated 2.27% a year on top.

🐻 Bear case. The core take rate is in structural decline: rebates grew +22% in FY23, +16% in FY24 and +16% in FY25 against gross growth that could not keep up, and +22% again in Q2 2026. Legislators on two continents are attacking the fee pool (interchange rulings in the UK; credit-routing legislation in the US, reported to have White House backing in October 2026), and stablecoins plus account-to-account rails aim precisely at cross-border, the most profitable flow. At 43× trailing earnings before interest and taxes (EBIT) on the terminal value check, the price leaves no room for the fade to begin.

1. The take rate only goes one way. Every renewal cycle transfers economics from the network to issuers. If the yield slide of ~1pp a year persists, gross volume must compound near double digits forever just to hold net revenue growth — against a US consumer base that is already card-saturated.
   Answer on the record: pricing is the offset — cross-border assessments grew +20% currency-neutral in Q2 2026 on +12% volume, and transaction processing assessments +12% on +9% switched. Pricing power is demonstrated, not asserted. The kill survives only if the Decisive checkpoint fails.
2. Regulation reprices the pool. A UK tribunal ruled multilateral interchange fees breached competition law (Jun 2025, reported); the November 2025 merchant settlement (reported) cuts average fees ~10bp for five years; US credit-routing legislation would let merchants steer transactions to cheaper networks.
   Answer: Mastercard does not keep interchange — issuers do. Its own assessments are the smaller, service-priced slice, and two decades of Durbin-era debit regulation coincided with the network compounding at the rates in the Series. The risk is real but historically absorbed; it is flagged, not fatal, on this record.
3. Bypass skips the toll road entirely. Stablecoins settle cross-border for fractions of a cent; account-to-account systems (and AI agents instructed to minimise cost) route around cards.
   Answer: management is buying and building on the new rails (BVNK; stablecoin settlement partnerships, reported) and ~60% of VAS attaches to services any rail still needs — fraud, identity, tokenisation. Bypass is a decade-scale migration, and the checkpoints measure cross-border growth while it plays out.

When you pay with a Mastercard, four parties touch the transaction: your bank (the issuer), the merchant's bank (the acquirer), the merchant, and Mastercard in the middle — setting the rules, switching the authorisation, and clearing settlement. Mastercard never lends you the money and never carries your balance; the issuer does. Mastercard charges assessments for the network itself: domestic assessments on same-country volume, cross-border assessments (the richest) on international volume, and transaction processing assessments per switched message. From those gross fees it pays rebates and incentives back to issuers and partners to win and keep portfolios — that payment, $20,522M in FY25, is subtracted before revenue is reported, which is why the net yield is the number that matters. On top of the network it sells services — fraud scoring, identity and authentication, tokenisation, consulting and data insights — as value-added services and solutions (VAS), now 41.6% of net revenue. The model needs almost no capital: capex was 2.1% of TTM revenue, and nearly all owner earnings are available to return — $11,727M of buybacks in FY25 alone.

Mastercard guides growth ranges (currency-neutral, excluding inorganic activity; Generally Accepted Accounting Principles (GAAP) and non-GAAP labelled on each call), not points. Across the twelve quarters read: FY2026 was guided at the high end of low double digits for net revenue on January 29, 2026, maintained on April 30, 2026, and on July 30, 2026 the range was held with delivery expected higher within it after a stronger first half. Q2 2026 itself was guided to the low end of low double digits and delivered +12% currency-neutral — above the guided end. Unfavourable items are named where they first appear: portfolio lapping and the Capital One debit migration were flagged through 2025–26 with the revenue weight placed in 2027, before they land. The full guided-vs-delivered ledger (append-only) is maintained in the dossier and cited by the grade.

At June 30, 2026 (filed Q2 10-Q): cash $11,291M; total debt $24,643M; net debt $13,352M — 0.8× TTM owner earnings, covered by under ten months of owner earnings. There is no lending book and no deposit base; settlement balances are short-cycle pass-throughs of the four-party model. Debt funds buybacks and acquisitions at investment-grade cost while the business itself consumes almost no capital (capex 2.1% of revenue). Equity of $5,611M is thin by design — decades of buybacks — and is not a solvency signal for a fee network with this cash conversion. Verdict: healthy, deliberately levered, no flag.

Method, briefly. Solve on enterprise value (EV) at a fixed 10% required return over 10 years with 2% terminal growth; terminal value = year-10 owner earnings × 12.75. The solved unit is TOTAL owner earnings — never earnings per share, whose growth would embed the buyback and hide the flat-vs-credited distinction. Base: TTM operating cash flow $17,437M − capex $735M = free cash flow (FCF) $16,702M − stock-based compensation (SBC) $615M = owner earnings $16,087M. No archetype switch: Mastercard has no lending book or float (see Gaps for the filed ruling summary), so cash flow is the right lens and no EPS-derived base is used.

EV bridge. Market cap $540,866M (876.01M shares × $617.42) + debt $24,643M − cash $11,291M = EV $554,218M (balance sheet at Jun 30, 2026, filed).

Buyback credit. The TOTAL bar (+15.7%/yr) is identical flat or credited — the shrink restates it per share. Credited per-share solve +18.4%/yr runs at the NAMED demonstrated input: DEMONSTRATED_SHRINK = 1 − 906/927 = +2.27%, computed in the governing script from the two filed diluted weighted-average share counts (FY2025: 906M; FY2024: 927M; FY2025 Form 10-K). $7.8B of authorisation remained at Jul 27, 2026 and Q2 2026 repurchases were $4.9B — the pace behind the credit is current. Rates above the demonstrated input are assumed by construction: at 3.0% shrink the per-share equivalent is +19.3%/yr; at 1.0%, +16.9%/yr. Timing record: repurchases rose every year FY21–FY25 through drawdowns and highs alike — dependable shrink, unproven price discipline (see Management).

Checks. Terminal value is 61.3% of EV, and the terminal value implies 43.1× TTM EBIT — far above a mature multiple, which is the model telling the truth about itself: the price only works if mid-teens compounding runs most of the decade. Competitive advantage period (CAP): at the delivered +17.2%/yr pace, the price is covered in ~9 years. Hurdle sensitivity (total bar): 9% → +13.5%/yr; 11% → +17.7%/yr. FCF memo: solving on raw FCF ($16,702M) instead of owner earnings would lower the bar by roughly 0.6pp — the SBC charge is kept, per house rule. What you must believe: net revenue compounding low-double-digits with operating margin held near 58%, the net-yield slide contained near its recent ~1pp/yr pace, and no legislative repricing of the assessment pool for ten years.

• Network effects & acceptance Answered — 3.7B cards, 200+ countries; both sides join because the other side is there.
• Switching costs Answered — issuer portfolios, tokenisation (over 40% of switched transactions) and VAS attach make migration a multi-year programme, as the Capital One debit exit itself shows.
• Pricing power Answered — assessment growth above volume growth in Q2 2026 (cross-border +20% currency-neutral on +12% volume).
• Net-yield defence Contested — pricing gains are being handed back through rebates at renewal; yield fell 51.0% → 48.7% in two years.
• Regulatory durability Flag — interchange rulings (UK), the reported merchant settlement, and US routing legislation all target the economics around the network.
• Bypass resistance Contested — BVNK and stablecoin settlement are hedges in progress; account-to-account migration is measured by the checkpoints, not yet by results.

B+  Michael Miebach, President & CEO since Jan 1, 2021 (joined 2010). CFO transition, pre-announced Jun 2, 2026: Sachin Mehra (CFO since 2019) moved to Chief Business Officer and Ling Hai (joined 2010; President, Asia Pacific, Europe, Middle East and Africa) became CFO on Aug 3, 2026 — an internal, explained succession, executed on schedule.

Guidance credibility A−. Twelve quarters of ranges met or beaten at the top end (ledger in the dossier; Q2 2026 delivered above its guided end). Execution A−. VAS built from $7.9B to $13.3B in three years while the network compounded; Recorded Future integrated into the security stack. Candor B+. Unfavourable drivers are named early (Capital One's 2027-weighted drag; a Venezuela-driven cross-border tailwind disclosed as such) — held off A-range because cohort-level rebate costs, the number that would prove yield discipline, are never disclosed. Capital allocation B+ — the cap. Buybacks of $5,904M → $11,727M (FY21→FY25) ran at any price, including $4.9B in Q2 2026 near highs; acquisitions are sized with discipline (Recorded Future ~$2.65B; BVNK up to ~$1.8B reported — each under a quarter of FCF) but the repurchase programme shows no price sensitivity, so the demonstrated 2.27% shrink is dependable while its value capture is unproven. Alignment B+. 2025 annual incentive funded at 145% of target and the 2023 performance share unit (PSU) cycle paid 131.8% (DEF 14A, Apr 27, 2026); say-on-pay drew 96% support at the 2025 meeting as reported in the 2026 proxy; Form 4 census to Oct 16, 2026 (115 filings over two years) shows routine plan settlements with no open-market purchase cluster.

Grade re-verified Oct 16, 2026 — proxy (DEF 14A, Apr 27, 2026), Form 4 census extended to Oct 16, 2026, buyback and share-count series from filed 10-K/10-Q, 12-quarter calls record. Verdict: B+ stands.

• Leverage — PASS. Net debt 0.8× TTM owner earnings; no lending book, no deposit flight risk.
• Moat — FLAG. Routing legislation or a fee-pool repricing would cut the assessment base the whole model compounds. Live and dated; measured by checkpoints 1 and 3.
• Management — PASS. No candour failure, accounting flag or succession disorder on the verified record.
• Bias — FLAG. A quality compounder at 96.8% of its 52-week range invites extrapolation: the desk's own bar (+15.7%/yr for a decade) is an extraordinary claim by base rates, however ordinary it looks beside the last four years.
• Concentration — PASS. No single customer, country or portfolio is large enough to break the model; the largest named exit (Capital One debit) is guided as a 2027-weighted, partial drag.

No veto fires. Two flags stand — moat (regulatory) and bias (extrapolation) — and both are wired to dated tests rather than left as vibes.

Grading convention: an outcome between a checkpoint's PASS and FAIL lines is MIXED — neither passed nor failed — and the item is re-graded at the following print.

Update protocol: this page is re-scored against the lines above after the October 29, 2026 print; every threshold was set on October 17, 2026, before the results, and will not be redrawn after the fact. Re-basing occurs only on a restatement, a change in the filed share count basis, or a corporate action — each recorded in Corrections.

First standard edition (October 17, 2026). No prior-edition corrections exist. One build-time restatement is recorded for transparency: FY2024 buybacks are printed at $10,954M, the FY2025 Form 10-K comparative, superseding the $11,035M first filed for FY2024 — a source-vintage correction made before publication, recorded here so the figure's provenance is never in doubt.

• Rebates and incentives FY22–FY24 are derived from filed growth rates (anchor FY25 filed); cohort-level rebate costs are never disclosed by the company.
• The FY21 payment-network / VAS split is not presented — the category presentation changed and the split was not extracted; FY22 onward is filed or derived as labelled.
• BVNK's price allocation and revenue contribution are not yet in a filed 10-Q at build date; acquisition terms are reported (Tier 3).
• The November 2025 merchant settlement's exact terms and the 2026 say-on-pay percentage sit in documents summarised at Tier 3 at build date; both are labelled "reported" where used.
• Archetype ruling (summary; full text in the dossier): fee-network, NOT float/lending — no cardholder lending book, settlement balances are short-cycle pass-throughs, rebates are contra-revenue, and free cash flow is real (capex 2.1% of revenue). The owner-earnings base therefore governs with no EPS-derived alternative.

• T1 · Filings: FY2025 Form 10-K (filed Feb 11, 2026; acc. 0001141391-26-000013) — revenue disaggregation, rebates anchor, WASO 906M/927M, buybacks; FY2021–FY2024 Forms 10-K — series; Q1 2026 Form 10-Q (acc. 0001141391-26-000031); Q2 2026 Form 10-Q (filed Jul 30, 2026; acc. 0001141391-26-000083) — headline quarter, Jun 30, 2026 balance sheet; DEF 14A (Apr 27, 2026; acc. 0001141391-26-000021) — pay, incentives, say-on-pay history; Form 4 census, EDGAR submissions index, CIK 1141391 (115 filings, Oct 17, 2024 – Oct 16, 2026).
• T2 · Transcripts & company materials: 12 earnings calls, Q3 2023 – Q2 2026 (Motley Fool full-text library; IR webcast archive cross-check for the two most recent); Q2 2026 and Q4 2025 earnings presentations (investor.mastercard.com); IR announcement of Oct 8, 2026 setting the Q3 2026 print for Oct 29, 2026.
• T3 · Reported (colour only, labelled on page): price history — StockAnalysis / Finnhub / Zacks (Oct 16, 2026 close $617.42); merchant settlement and UK tribunal summaries; Credit Card Competition Act status and reported White House backing (Oct 2026); BVNK closing reports. No analyst estimates used as inputs anywhere.