Is American Express a Good Long-Term Investment? AXP Growth, Profitability and Risks

American Express did two things before the US open on Friday 23 October that are not normally announced on the same morning. It raised full-year 2026 earnings guidance to $18.30-$18.60 a share on a reported basis, up from $17.30-$17.90, and it took a $336m pretax restructuring charge to eliminate roughly 5% of its positions worldwide — about 3,500 roles out of a workforce of some 78,000, with the chief financial officer, Christophe Le Caillec, tying the programme explicitly to automation and to stripping out layers of management. Third-quarter revenue was a record $20.1bn, up 9%, and net income rose 28% to $3.71bn. The shares fell 3.68% to close at $313.22.

Anyone asking whether American Express is a good long-term investment now has an unusually clean test case, because the release contains the bull and bear arguments side by side. What the numbers support is a structural margin story: a one-off $336m spent to automate work out of a business whose revenue growth guidance for the year is unchanged at about 10%, with credit costs falling and card balances still compounding at 8%. What would have to go right is the top line. Billed business growth slowed to 8% in the quarter from 9% across the first half, commercial spending grew just 3%, and revenue came in fractionally below the roughly $20.12bn the market expected. A cost programme creates long-term value only if volumes hold while the headcount falls; on this quarter's evidence the cost side is moving faster than the revenue side, and that is the tension to resolve rather than a question the print answers.

What American Express reported

Start with the two earnings figures, because the distance between them is where most of the misreading of this quarter will happen. Reported diluted earnings were $5.42 a share. Adjusted diluted earnings were $4.68, up 13% year on year, against a consensus near $4.48. The $0.74 difference is not noise: it is the net effect of two large one-off items that the company excludes from its adjusted measure — a gain on the sale of its equity stake in Global Business Travel Group, worth $1.45 a share pretax and $1.11 after tax, less the restructuring charge at $0.49 a share pretax and $0.37 after tax. Netted, those items contributed $0.74 a share that the adjusted number strips out.

What the adjusted figure does not strip out is worth as much attention. Inside the $4.68 sits an increase in legal reserves of $0.51 a share, which the company says includes the portion of the Office of the Comptroller of the Currency civil money penalty that had not previously been reserved for, plus a reserve for an anticipated settlement of an unrelated matter. Also inside it: a $211m reserve release driven by CECL model enhancements, which is the main reason the provision line fell as far as it did. In other words the adjusted beat both absorbed a half-dollar of fresh legal provisioning and leaned on a one-off credit reserve release. The underlying operating result is closer to consensus than the headline beat implies in either direction.

The volume lines were the softer half of the release. Billed business rose 8% to $454bn, a step down from 9% in the first half, with roughly 60 basis points of the deceleration attributable to co-brand portfolio exits. Commercial spending grew 3%, or 5% excluding those exits, with the middle market segment cited as underperforming on competitive pressure. International billed business growth moderated to 11% from 13%, which management framed as a twenty-second consecutive quarter of double-digit growth — accurate, and also a decelerating series. Card member balances grew 8%, broadly in line with billed business.

Credit was the clear positive. The provision for credit losses fell 23% to about $985m, reflecting a reserve release this period against a build in the prior year, and the net write-off rate improved to 1.8% from 1.9% a year earlier. Business development expense rose 11% on higher partner payments and client incentives. Revenue growth guidance of about 10% for the full year was reaffirmed, and the adjusted earnings range was set at $17.60-$17.90 alongside the $18.30-$18.60 reported range.

Why the market reacted the way it did

A 3.68% fall on a raise and a beat looks contrary until you separate the two announcements. The guidance raise is largely mechanical: a $982m pretax gain on the GBTG disposal flows through reported earnings, so the reported range had to move up. The adjusted range of $17.60-$17.90 is the figure that reflects the operating business, and it is a far more modest revision. Investors who had been reading the headline $18.30-$18.60 as an upgrade to earning power were correcting a category error.

The restructuring was read the same way — not as a cost saving to be capitalised, but as a signal about the revenue outlook. Companies rarely remove 5% of their people when volumes are accelerating. Cutting 3,500 roles in the same release that shows billed business decelerating and commercial spending at 3% invites the interpretation that management is sizing the cost base to a slower growth environment. That is a reasonable inference and an unprovable one, which is precisely why the stock moved on it.

There is a third strand. American Express is a premium consumer franchise whose valuation has rested on the proposition that its cardholder base is insulated from the spending slowdown affecting the mass market. Billed business at 8% rather than 9% is not a collapse, but it is the first quarter in a while where that insulation looks a little thinner, and it arrived with a middle market commercial business under visible competitive pressure.

The Openbook read

The five-factor framework is where this quarter resolves into something usable. Current factor detail sits on the AXP equity page, and the Openbook screener shows how the payments and consumer finance names compare on the same measures.

Momentum

Weakened but not broken. A 3.68% fall on results day is a normal-sized reaction to a mixed print rather than a trend reversal, and it followed a prior close of $325.17. The more meaningful momentum signal is in the operating series rather than the share price: billed business decelerating from 9% to 8%, international from 13% to 11%, commercial at 3%. Three decelerating volume series is a momentum observation about the business, and it is the honest reason to mark this factor down.

Growth

Solid in absolute terms, slowing at the margin. Revenue of $20.1bn, up 9%, with about 10% reaffirmed for the full year, is a good number for a business of this maturity. Card balances compounding at 8% and net card fees growing in the mid-to-high teens point to a franchise still adding economic value per customer. But growth guidance reaffirmed is not growth guidance raised, and the quarter's revenue came in marginally below expectation. This is a steady-growth score, not an accelerating one.

Profitability

The strongest factor, and the one the restructuring is aimed at. Net income of $3.71bn on $20.1bn of revenue, up 28%, is a high return even allowing for the disposal gain. The $336m charge is the interesting entry: spent once, against a cost base it is intended to permanently reduce, in a business where the revenue guidance did not change. If the programme delivers and volumes hold, the arithmetic is straightforwardly accretive to margin. The caveat is that the same release carried $0.51 a share of legal reserves and an 11% rise in business development expense, so not every cost line is moving downward.

Solvency

Comfortable, and improved by the credit data. A 23% fall in the provision for credit losses to roughly $985m and a net write-off rate of 1.8%, down from 1.9%, describe a loan book performing better than it was a year ago, which matters more for Amex than for a pure network because it carries its own credit risk. Nothing in this quarter raises a balance sheet question; the reserve release that helped earnings is itself a statement about asset quality.

Reward/Risk

Finely balanced, and the balance depends on one variable. The reward case is a high-return franchise executing a permanent cost reduction with credit costs falling. The risk case is that the cost programme is a response to decelerating volumes rather than an independent efficiency gain — in which case the margin benefit arrives alongside the revenue shortfall that prompted it, and the two largely cancel. Distinguishing between those requires at least two more quarters of billed business data. Until then the reward/risk score should be read as wide rather than favourable.

The read-across

For the pure networks, Visa and Mastercard, the relevant datapoint is the volume deceleration rather than the job cuts. Amex's billed business is skewed to affluent consumers and corporate travel and entertainment, so an 8% print with commercial at 3% is an early signal on the discretionary and business spending that feeds the networks' own volume lines. The networks do not carry credit risk, so Amex's reserve release reads across to them barely at all; its spending data reads across directly.

For lenders with overlapping consumer exposure, including Capital One, the useful read is the opposite way round. A 23% fall in provisions and a 1.8% write-off rate at the premium end of the credit spectrum is a constructive signal for consumer credit quality broadly, even if Amex's cardholder base is not representative of subprime or near-prime books.

The wider read-across is to the labour line across financial services. Amex has now put a number on what automating management layers is worth to it — $336m to remove 5% of positions — and framed it as preparation for AI rather than as retrenchment. Peers will be asked the same question on their own calls, and the framing American Express has chosen makes it harder for them to present a comparable programme as anything else.

What to watch next

The fourth-quarter and full-year results are the next scheduled test, and they carry two things that matter more than the quarter itself: the first 2027 guidance, and the first opportunity to see whether billed business stabilises at 8% or steps down again. Revenue growth of about 10% for 2026 has been reaffirmed rather than raised, so the full-year outcome will show how much of that depended on the first half.

Three specific items are worth tracking. First, the restructuring's cash and timing profile — the company has disclosed the $336m charge and the approximate 5% but has not published a completion timetable or a quantified annual saving, and that saving is the entire basis of the margin argument. Second, whether commercial spending recovers from 3%; the middle market weakness is the part of the franchise most exposed to competition rather than to the cycle. Third, the legal reserve line, which absorbed $0.51 a share this quarter including the balance of the OCC civil money penalty and a provision for an anticipated settlement of a separate matter — whether that proves to be a clearing of the decks or the start of a run of similar charges will change how much of the adjusted earnings number can be treated as repeatable.