American Express and the Toll Booth’s Asking Price
What the business is
American Express is a toll bridge that issues the pass, operates the crossing, and sometimes lends travelers the fare. The arrangement is economically attractive, though rather less simple than an ordinary payment network.
Its closed-loop model combines card issuance with payment processing. The company earns merchant discount revenue, net card fees, and net interest income from card loans and revolving balances. Premium customers tend to spend more, helping American Express charge merchants higher rates than Visa or Mastercard. Unlike a network that merely routes transactions, however, American Express also keeps credit risk on its balance sheet.
The deeper business is a premium consumer brand attached to a membership system. Net card fees reached about $10 billion in 2025, increasing 18% from the prior year and extending 30 consecutive quarters of double-digit growth. New customers are also arriving through paid products: 73% of new accounts came from fee-paying cards, while Gen-Z spending in the US consumer segment grew 38%.
I see three mutually reinforcing sources of advantage. The first is a two-sided network linking affluent cardholders with merchants. The second is a brand associated with service, access, and status. The third is switching friction created by accumulated rewards, credits, lounges, and integration with corporate expense systems.
The moat appears to be widening, although the guards are becoming more expensive. American Express had 127.6 million cards, up 8%, and its recurring fee stream is becoming more important. Its closed-loop structure also provides proprietary data from both sides of a transaction, supporting fraud detection, credit decisions, risk management, and targeted offers. AI models may become commonplace; proprietary transaction histories are less generous.
Competition remains the counterweight. Capital One and Discover form another integrated-network rival, while Chase Sapphire and Robinhood are contesting the premium-card market. Rewards and member benefits can reinforce loyalty, but escalating benefit costs could allow customers to enjoy the moat more than shareholders do.
What the numbers say
Annual owner earnings are $10.185 billion, close to 2025 net income of $10.1 billion. The thesis treats net income as a reasonable approximation of owner earnings because capital expenditure is relatively light for this financial business.
The operating record supports the quality argument. Five-year average ROE was 32.5%, with a standard deviation of 2.6 percentage points. High returns are useful; high and stable returns are better evidence that retained earnings have been deployed productively.
Capital allocation has emphasized organic growth and shareholder returns rather than large acquisitions. American Express returned about $7.6 billion to shareholders in 2025, and the quarterly dividend increased from $0.82 to $0.95, a rise of 17%. Repurchases are the less convincing part of the record because they have resembled a standing program rather than a response to valuation. Regularity is admirable in dental care, less so in buying expensive shares.
- Market cap (price-adjusted, 2026-10-09 close): $207.44 billion
- Owner-earnings yield (price-adjusted): 4.91%
- Share price (2026-10-09 close): $308.17
- 10-year Treasury: 5.24%
The central comparison is plain: a 4.91% owner-earnings yield against a 5.24% Treasury. The spread is negative 0.33 percentage points. Equity owners accept business, credit, competitive, and regulatory risk, yet the current earnings yield is lower than the government-bond rate.
The evidence’s parsed intrinsic-value reference is $122.0 billion. The price-adjusted market cap carries a 70% premium to that figure. American Express deserves a better valuation than an ordinary business, but quality does not turn every asking price into value.
What Wall Street believes vs what I see
The available headlines are constructive. They describe American Express as cheaper than peers, more resilient than it appears, and supported by a strong business model. I agree with much of the underlying praise. The premium brand, recurring card fees, network effects, proprietary data, and stable historical ROE form a substantial franchise.
My disagreement begins at the cash register. “Cheaper than peers” is relative; owner earnings are absolute. A 4.91% owner-earnings yield against a 5.24% Treasury does not become generous merely because another security is dearer. Selecting the least expensive bottle on a costly wine list does not make it tap water.
The risks need not become disasters for the valuation to prove demanding. A premium-card benefits war could cause rewards and service costs to outrun fee and discount-revenue growth. Credit deterioration would expose the lending side of the model, while merchant-fee regulation could pressure a central revenue source. Annual-fee fatigue could also test customer renewals.
I therefore see a strong business whose membership economics and data advantage remain attractive, but whose price offers little compensation for those uncertainties. The market appears to understand the quality quite well. It may even be admiring it with both hands.
Verdict
Verdict: pass.
The 4.91% owner-earnings yield does not exceed the 5.24% Treasury, so the shares fail the rubric’s watch threshold. The $207.44 billion market cap also stands at a 70% premium to the $122.0 billion intrinsic-value reference, providing no evident margin of safety. The verdict would improve to watch if the price-adjusted owner-earnings yield rose above the Treasury; circle would additionally require the yield hurdle and market-cap discount specified by the rubric.
This analysis is AI-generated, educational, and not investment advice. Figures may contain errors or be delayed. Disclaimer