Walmart's $2.9bn Tariff Refund Lands (20 August 2026): A 7-Cent EPS Beat, a 90bp Comp Miss, and a 9% Fall

Walmart reported second-quarter fiscal 2027 results before the US open on Thursday 20 August 2026, and the tariff refund it had deliberately excluded from guidance arrived roughly 20% larger than management had sized it: approximately $2.9 billion, not $2.4 billion. Revenue was $187.9 billion against about $186.8 billion expected, adjusted EPS was $0.81 against $0.74, and the full-year outlook was raised for the first time since February. The stock closed about 9% lower. The reason is the whole point of this report: the money came in through gross margin, while Walmart US comparable sales grew 2.6% against roughly 3.5% expected — and once you strip out eCommerce, the physical store contribution to comp was negative.

This piece previewed the quarter on 16 August on a single argument: that a refund of duties already paid is a cash event with a legal timetable, that earnings guidance is a forecast about an operating business, and that the report would put the headline number and the most important disclosure in different sections. That is what happened, and the 9% fall is the market performing the separation in public.

What actually happened

The scoreboard, against the company's own May guidance and the street:

Revenue and earnings expectations come from an LSEG survey and the comp expectation from FactSet, both as reported by CNBC; every company figure here is from Walmart's own second-quarter release.

Underneath, the growth engines that have carried this business for two years all worked. Global eCommerce grew 23%, with Walmart US eCommerce up 24% on strength in store-fulfilled delivery. Global advertising grew 38%, with Walmart Connect up 43% excluding VIZIO. Membership fee revenue grew 17% globally, and Walmart+ net additions hit a record second-quarter high. Walmart International net sales rose 12.8% to $35.2 billion, 7.9% in constant currency. Sam's Club US comps ex-fuel grew 4.4% on transactions up 7.0% — with average ticket down 2.5%, which is what trade-down looks like in the data.

The refund went into the shelf, not the earnings line

Walmart never printed "$2.9 billion" as a line item. It printed a gross profit rate. The consolidated rate rose 96 basis points to 25.4%, and the release attributes that "primarily" to tariff refunds, partially offset by price investments and higher fuel costs in distribution and fulfilment. To see how large a 96 basis point move is for this company, put it next to the four quarters before it.

Within Walmart US the rate rose 158 basis points to 29.4% on gross profit of $36.8 billion. At Sam's Club US, operating income of $678 million was up 44.3%, with the release citing tariff refund benefits and membership growth.

Here is the part that matters for anyone reading the beat as strength. Because the refund arrived inside gross margin rather than as a non-GAAP adjustment, it is inside the $0.81. Walmart said so in as many words: setting aside the net impact of refunds and the price investments funded by them, underlying operating income growth was at the top end of the guidance range. Top end of the range is a good quarter. It is not a 17.4% adjusted operating income increase, and it is not a seven-cent beat.

Why a 9% fall on a beat is not irrational

Walmart US comparable sales grew 2.6% excluding fuel, against 4.6% a year earlier and roughly 3.5% expected. CNN puts that as the slowest since the February-to-April window of 2020.

The composition is worse than the number. Walmart discloses how much of its comp came from eCommerce: approximately 510 basis points this quarter, against approximately 420 basis points of the prior year's 4.6%. Do the subtraction. Everything that is not eCommerce contributed roughly -2.5 percentage points to comp this quarter, against roughly +0.4 a year ago. Transactions rose 1.5% and average ticket 1.1%, so the customer count is still growing — but the store, as a channel, went backwards while digital did all the work and then covered a hole.

Some of that is identifiable and non-recurring. The release cites an 80 basis point headwind to comp from health and wellness, and the Walmart US segment detail attributes 125 basis points to pharmacy deflation from a new maximum fair price regulation effective 1 January — a rule change that cuts revenue per prescription without cutting the number of prescriptions filled. Strip that out and the picture improves materially. It does not become a 3.5% comp.

The forward guide compounded it. For the third quarter Walmart guided net sales growth of 3.0% to 3.75% and adjusted EPS of $0.62 to $0.64, against a Q3 FY26 base of $0.62 — roughly flat to plus 3%, immediately after a quarter that grew adjusted EPS 19%. Part of that is mechanical: management flagged a headwind of over 100 basis points to Q3 sales growth from a timing shift of Flipkart's Big Billion Days between the third and fourth quarters. The larger part is deliberate. Chief financial officer John David Rainey said the operating income outlook "reflects the continued prioritization of tariff refunds received in Q2 into customer experience and price investments in the second half," and asked investors to "consider Q2 and Q3 performance together to assess the underlying growth of the business." A company that tells you to average two quarters is telling you one of them is flattered and the other depressed by the same decision.

Full-year guidance did rise — net sales growth of 4.0% to 5.0% in constant currency, adjusted operating income growth of 7.0% to 8.5%, adjusted EPS of $2.80 to $2.87, capital expenditure to approximately 4.0% of net sales from 3.5%. Note the last one: the raise came with more spending, and the top of the new EPS range still sits below the roughly $2.89 the street had been carrying into the print.

Same ruling, four different answers

The most useful control experiment is the rest of the sector, reporting the same week off the same legal event.

Refund amounts other than Walmart's are as reported by CNN; Target's breakdown is from its own second-quarter release.

Target's treatment is the clean counterexample. It named the number, showed exactly what it did to EPS and margin, and carried approximately $1.65 of it into the full-year range. A reader can back it out in one subtraction. Walmart's is harder to isolate by design, because a meaningful share of it was already converted into lower shelf prices before it reached the income statement. Neither approach is wrong. But they produce very different-looking quarters from an identical cause, and any comparison of retail margins this season that ignores the distinction is comparing accounting choices rather than businesses. We set out the same separation problem before the print in the Home Depot Q2 piece.

The scale behind all of this: the Supreme Court's February decision invalidated tariffs imposed under the International Emergency Economic Powers Act, and refunds began flowing in May. CNN, citing a Customs and Border Protection court filing, reports that $168 billion was collected from 330,000 importers and that $100 billion had been sent out as of 31 July. The surviving tariff layers are unaffected, including the Section 232 measures we covered in the drone tariff explainer.

The offsetting force nobody guided for: fuel

The refund is disinflationary at the shelf. Something else in this report is not.

Walmart told investors it expects more than $2 billion of added costs this year from higher fuel prices, and higher fuel costs in distribution and fulfilment are cited repeatedly in the release as a partial offset to the gross margin benefit. Rainey described "arguably a softer consumer environment than in February," and said on the earnings call that when petrol prices go above $4 a gallon "there's a psychological impact to that. That there are choices that consumers are making" (CNN). Chief executive John Furner framed the price response directly: "Customers tell us they're still feeling some pressure," and said the best prices across a basket of goods help build trust "at a time when many households are carefully managing their budgets" (NBC News).

That cost line traces back to the energy complex we have been following: crude and, more acutely, distillate. Diesel is the fuel that moves freight, and its crack spread has been at records — we covered the mechanism in the diesel crack spread piece and the supply side in the Strait of Hormuz analysis. A retailer absorbing a fuel bill of that size is the downstream end of the same story.

So the honest summary of Walmart's contribution to the inflation picture is two forces pointed in opposite directions: a finite, one-off refund being converted into durable shelf-price cuts on imported goods, against a recurring energy cost that is currently rising. Neither cancels the other cleanly, and the goods and energy components of CPI will settle it, not a management quote.

Where this actually reaches an instrument

The chain is: retailer pricing decisions → goods prices → the goods component of core inflation → the Fed's reaction function → the rate factor in the dollar. Each arrow is real and each one attenuates the signal.

The refund cuts across it twice. It lowers goods prices where it is spent, which is disinflationary; and it does so with money the government has returned, which is fiscal rather than monetary. But it is one-time. When the claims run out, so does the funding for those price cuts, and the base effect reverses. A disinflationary impulse with a known expiry date is not the same input as a structural one — which is one reason a single quarter of retail margin commentary should not move a rate expectation, and did not.

What it does move is how the consumer complex is priced. A comp miss at the largest US retailer, arriving with an explicit statement that the consumer is softer than in February, is a read on volumes rather than on policy — faster and shallower in its transmission to an equity index than to a currency. For the dollar specifically, the durable factors are still the ones on the USD page, and we traced how rate expectations moved through real yields in the real-yield and gold piece.

Which scenario landed

The preview mapped four outcomes. Two happened at once, which is itself the lesson.

The row the preview did not have was the comp. Expectations for the operating business had run ahead of it, the refund had nothing to do with it, and that is what the 9% priced. Separating the cash event from the operating business was the right discipline going in; the market applied it within minutes of the release, and it is a reasonable prior for the next few quarters of retail reporting, because every one of these companies now has a one-off refund somewhere in its margin line and a choice about whether to tell you where.

Read more about how this site frames scheduled events on the about page, and see where the five factors currently score the dollar on the USD page.

Educational macro context only — not investment advice.