Walmart Q2 Preview (20 August 2026): The $2.4bn Tariff Refund That Is Deliberately Not in Guidance

Walmart reports second-quarter fiscal 2027 results before the US open on Thursday 20 August 2026, at 6 a.m. CDT, with the call at 7 a.m. CDT. The company has already told the market what it expects: net sales up 4.0% to 5.0% in constant currency, adjusted operating income up 7.0% to 10.0%, and adjusted EPS of $0.72 to $0.74 against $0.68 a year ago. Consensus sits at $0.74 — the top of that range, and no higher. The number that is not in any of those figures is roughly $2.4 billion of tariffs Walmart paid under the International Emergency Economic Powers Act and may now recover, which the Supreme Court invalidated in February. Management did not merely omit it; it said so out loud, and explained why.

That omission is the most instructive thing on the calendar this week, because it separates two ideas that get collapsed together every earnings season. A refund of duties already paid is a cash event with a legal timetable. Earnings guidance is a forecast about an operating business. Walmart drew the line between them explicitly, and in doing so created a report where the headline EPS number and the most important disclosure are likely to be in different sections.

What the company has already committed to

Most earnings previews have to reconstruct expectations from analyst notes. This one does not, because Walmart published a guidance table on 21 May 2026 alongside first-quarter results, including the prior-year base it is measuring against. That base matters: net sales of $175.8 billion is not the same as the $177.4 billion of total revenue reported for that quarter, because revenue includes membership and other income. Growth rates quoted against the wrong denominator are the single most common error in reading this company.

The full-year picture is where the pressure sits. Walmart's fiscal 2027 outlook — net sales up 3.5% to 4.5% in constant currency, adjusted operating income up 6.0% to 8.0%, adjusted EPS of $2.75 to $2.85, capital expenditure around 3.5% of net sales — was set on 19 February 2026 and left completely unchanged in May. Street consensus for the year has drifted to roughly $2.89, above the top of that range. That gap is not a scandal; it is normal for a company with a long history of guiding conservatively. But it does define what the 20 August report has to do. Meeting the quarter is not sufficient to validate the street's full-year number. Something in the second half has to be revised.

The $2.4 billion that is deliberately absent

On Friday 20 February 2026, the Supreme Court ruled 6-3 that the International Emergency Economic Powers Act does not grant the President authority to impose tariffs of indefinite scope, invalidating both the reciprocal tariffs and the trafficking-related tariffs levied under that statute. Because those duties were paid by importers of record at the border, the ruling created a refund claim rather than merely a prospective rate change. Estimates of total IEEPA collections run to roughly $175–179 billion (Penn Wharton Budget Model), and US Customs and Border Protection has since built a phased administrative process through which importers of record, or brokers authorised by them, file claims.

Walmart is one of those importers of record. Its Form 10-Q for the quarter ended 30 April 2026 states that the company "is participating in the process established by the U.S. Customs and Border Protection for refunds of tariffs that the Company paid as the importer of record" under IEEPA, that "the timing, amounts and ultimate resolution of any refunds remain uncertain and subject to ongoing legal and administrative developments," and that "accordingly, the Company did not recognize any amounts related to these claims in the three months ended April 30, 2026."

The guidance table carries the same fact in a single line of small print: the outlook "does not assume any impact from IEEPA tariff refunds." And in commentary around the results, Rainey sized it — roughly $2.4 billion, less than half of 1% of US annual sales — and explained the treatment directly: "We felt it best to provide guidance that reflects our expectations for the underlying business, excluding any recovery of tariffs paid."

Why a refund is not automatically a beat

Three independent conditions have to be satisfied before a court ruling becomes an earnings number. The first is legal resolution: a claim in an administrative queue is not money. The second is recognition — accounting treatment does not permit an uncertain, unquantified recovery to be booked, which is precisely the reasoning the 10-Q sets out and precisely why the first quarter carried nothing. The third is discretionary, and it is the one most likely to surprise readers who treat the refund as found profit. Walmart has already said the money is intended for price: Rainey framed price investment as the best available use of capital in the current environment and said the company would "definitely bias and try to prioritize" directing refunds toward lower prices, citing consumer pressure from fuel costs.

Follow that through the income statement. A refund that lands in the quarter and stays there is a one-off benefit. A refund that funds price cuts shows up as lower prices on the shelf, which mechanically compresses the gross margin rate, which is then offset — partially, wholly, or more than wholly — by higher unit volumes and by the higher-margin businesses that ride on those volumes. Walmart's advertising business grew 37% globally last quarter and membership fee revenue grew 17.4%; both scale with traffic. That is the actual argument for spending a refund rather than banking it, and it is a multi-quarter argument, not a one-print argument.

So the disclosure to look for on 20 August is not a number in the EPS line. It is whether management quantifies progress on the claims, whether anything has become estimable enough to recognise, and what it says about price investment in the second half.

What the ruling did not do

One layer of the tariff stack was removed retroactively. The rest was not. The February decision addressed IEEPA and the tariffs imposed under it; duties levied under other statutory authorities were untouched and continue to be extended to new categories — the 100% drone tariffs due to take effect on 3 September 2026 are a Section 232 action, and we set out that mechanism in detail in the Section 232 drone tariff explainer.

For Walmart specifically, the 10-Q notes that less than one third of what it sells in the US is imported, with most imports coming from China, Vietnam, Mexico, India and Canada. That is a smaller import exposure than the discourse usually assumes, and it is still live under the surviving architecture. A company recovering past duties while continuing to pay current ones has no reason to raise a full-year forecast on the strength of the refund alone — which is exactly what Walmart did not do in May.

The consumer walking into the print

The macro backdrop moved in the days before the report. The Census Bureau's advance estimate, released 14 August 2026, put July retail and food services sales at $763.6 billion, down 0.6% (±0.4%) on the month but up 5.0% (±0.5%) on the year, with the May-through-July period up 6.3% against a year earlier (US Census Bureau, CB26-131). Excluding autos the monthly change was -0.2%; excluding autos and gas, -0.3%.

Read that carefully, because the monthly and annual figures point in different directions and both are true. A 0.6% monthly decline against a 5.0% annual gain describes a consumer still spending more than a year ago but no longer accelerating. For a retailer whose growth last quarter came from transactions rather than ticket, the annual figure is the more relevant one — but the monthly figure is what moved rate expectations.

The week also stacks three consumer reports in sequence: Home Depot on Tuesday 18 August, Target on Wednesday 19 August, Walmart on Thursday 20 August. That ordering is useful rather than incidental, because the three sit at different points on the discretionary spectrum. We covered the housing-linked read in the Home Depot Q2 preview. Consensus for Target's quarter sits near $2.32 a share on revenue of about $26.12 billion. If a single narrative about the consumer holds across all three, it is worth more than any one of them; if the staples-weighted business reads differently from the discretionary ones, that divergence is itself the information.

The channel to the instruments this actually touches

The honest chain runs: 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.

Where that chain currently stands is measurable. Markets price a September hold as the base case, with hike odds just above 30% after a July jobs report that showed payrolls falling 23,000 and a benign July inflation print, against roughly 63% odds of at least one hike by year-end (CNBC). We traced how that repricing moved through real yields in the September hike odds and gold piece.

Against that, one retailer's margin commentary is a small input — but it is not zero, and the reason is aggregation. When the largest US retailer by revenue signals that it intends to route a windfall into shelf prices, its competitors face the same shelf. What would eventually show up in inflation data is not Walmart's decision in isolation but whether it forces a broader reset in goods pricing over two or three quarters — a thing to watch in the CPI goods series, not in Thursday's stock reaction. For equity index exposure the transmission is faster and shallower: the read on the consumer conditions how the whole consumer complex is priced into the autumn, but an index move on one guidance revision is sentiment repricing, not a change in the inflation path.

The scenario map

Not predictions — a map of what each outcome would mean, and what would have to be true for it.

The unifying discipline across all four rows: separate the cash event from the operating business, exactly as the company itself has done. A report that is read as one thing when it is really two is how a mechanical beat gets mistaken for an improvement, or a deliberate margin investment gets mistaken for a miss.

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.