US Retail Sales — July 2026

July produced both a 'biggest drop in a year' headline and a +5.0% year-over-year figure — the same report, different comparison windows. The advance estimate's ±0.4 percentage-point sampling error means the -0.6% itself is measured with meaningful uncertainty.^[1][2][3]

The single most consequential detail about this release is stated four times in the Census Bureau's own text: the figures are adjusted for seasonal variation and for holiday and trading-day differences, «but not for price changes». Retail sales are nominal dollars.^[1]

That matters most for the year-over-year comparison. A gain of 5.0% over July 2025 is not a measure of how much more people bought — it is how many more dollars they spent, inflation included. The Census Bureau warns about exactly this reading, noting the data «can be misleading if you do not acknowledge the role inflation (or any price change) plays», and that price change «could even be in the opposite direction» from the underlying trend.^[1][5]

Working the adjustment through changes the monthly picture too. KPMG calculated that real retail sales fell 0.7% in July once consumer prices are accounted for, against the nominal −0.6%. The Census Bureau does not publish a real series itself; inflation-adjusted consumption is produced separately by the Bureau of Economic Analysis.^[4]

The July figure is an advance estimate, drawn from «approximately 4,800 retail and food services firms» — a subsample of the larger monthly survey of about 13,000 units. One month later the advance figure is replaced by the estimate from that fuller sample, and it is revised again as annual benchmarks arrive.^[1][5]

The revisions are smaller than the three-stage structure suggests: the Census Bureau puts the average absolute revision to the advance estimate at about two-tenths of a percentage point. The more accurate expectation is not that headlines get rewritten, but that dollar levels shift quietly while the monthly percentage moves at the decimal. June 2026 illustrates it — the level went from $768.6 billion to $768.1 billion while the May-to-June change stayed at +0.2%.^[1]

So July's −0.6% should be read as a first estimate rather than a settled number. It has not yet been revised at all; the first revision arrives with the August release.^[1]

Market commentary on retail sales usually leads with the control group — sales excluding autos, gasoline, building materials and food services. It is worth knowing that this series does not appear in the Census release at all. The Bureau publishes three exclusion measures: excluding motor vehicles and parts, excluding gasoline stations, and excluding both.^[1]

The control group is computed by analysts from the published detail, and definitions vary slightly between houses. What gives it standing is the downstream use: the Bureau of Economic Analysis estimates most goods categories of personal consumption using a retail control method, so this subset feeds the consumption side of GDP.^[4]

The practical reason analysts prefer it is visible in July. The headline was dragged down by motor vehicles and by gasoline stations, where falling prices reduce dollar sales regardless of volume — and gasoline dollars move with prices in a series that is not price-adjusted. Stripping those out gave a smaller decline than the headline.^[1][4]

August 2026 advance retail sales are scheduled for September 16, 2026 at 8:30 a.m. EDT. That release will also carry the first revision to the July figure discussed here.^[1]

A separate and larger revision follows. The Census Bureau has published an «Intention to Revise» notice stating that monthly retail sales estimates are expected to be revised for historical corrections and the results of the 2023 and 2024 Annual Integrated Economic Survey, tentatively on September 28, 2026. Comparisons drawn from figures published before that date may not match the series afterwards.^[1]

July was revised. The Census Bureau states that the June-to-July change was "revised from down 0.6 percent (±0.4 percent) to down 0.5 percent (±0.2 percent)" — so the -0.6% quoted on this page was the advance estimate, and the current figure is -0.5% with July sales at $764.5 billion. This is exactly the revision behaviour described further down this page.^[6]

August 2026 advance sales were $773.9 billion, up 1.2 percent (±0.4 percent) from July and up 6.0 percent (±0.5 percent) from August 2025. The June-August period was also up 6.0 percent (±0.5 percent) year over year. So the single soft month did not continue — but note that these figures remain not adjusted for price changes.^[6]

The next release, covering September 2026, is scheduled for 15 October 2026 at 8:30 a.m. EDT.^[6]