U.S. September retail sales rise 0.8%, with gasoline stations accounting for seven-tenths of the increase

U.S. retail and food services sales continued to increase in September, but the most revealing feature of the report is where that growth occurred. Advance estimates released by the Census Bureau on October 15 put sales at $745.0 billion, up 0.8% from August and 6.3% from September 2025. The figures adjust for seasonal patterns and holiday and trading-day differences, but not price changes. The sampling margin of error is ±0.4 percentage points for the monthly change and ±0.5 points for the annual change, both at the 90% confidence level.¹

Gasoline-station sales rose 6.4% during the month, while sales excluding gasoline stations increased just 0.2%. Assessing whether demand strengthened broadly therefore requires looking at dollar amounts, industry composition and prices together. Industry composition determines where the growth signal is relevant; prices and quantities determine whether it reflects more active purchasing.¹

Archive photo: endcap displays at a Walmart in Wallkill, New York, on April 3, 2017. Photo: Daniel Case, “Endcaps at a Wal-Mart” / Wikimedia Commons; CC BY-SA 3.0.

Gasoline stations account for most of the dollar increase

Table 1 shows total retail and food services sales rising from $739.322 billion in August to $745.021 billion in September, an increase of $5.699 billion. Gasoline-station sales rose from $63.950 billion to $68.021 billion, an increase of $4.071 billion. Dividing those two published changes puts gasoline stations' share of the overall monthly dollar increase at approximately 71.4%. This is an arithmetic decomposition of official figures, not a causal attribution published by the Census Bureau, and it has no separately calculated confidence interval.¹

The other industries combined still expanded: sales excluding gasoline stations increased from $675.372 billion to $677.000 billion, or about 0.2%. Excluding only motor vehicle and parts dealers, sales rose 0.9%; excluding both those dealers and gasoline stations, growth returned to 0.2%. Motor vehicle and parts dealers themselves recorded a 0.2% increase. The 0.9% measure excluding autos still includes the sharp gain at gasoline stations. Excluding both categories gives a more useful view of the remaining retail activity.¹

Gasoline-station revenue must also be distinguished from gasoline volume. The survey classifies establishments by their main business, so sales at a type of store do not measure nationwide sales of a single product. Prices, quantities and product mix can all affect gasoline-station revenue; the report does not separate those effects. The 71.4% calculation cannot be recast as evidence that fuel prices explained 71.4% of retail growth.²

Archive photo: the entrance to Ourisman Volvo and Volkswagen in Bethesda, Maryland, on May 26, 2024. Photo: G. Edward Johnson, “Ourisman Volvo and VW Bethesda MD 2024-05-26 17-39-20 1” / Wikimedia Commons; CC BY 4.0. Original copyright notice: Copyright 2024, G. Edward Johnson. CC-By Attribution.

Industries outside gasoline stations diverge

The seasonally adjusted monthly point estimates show increases of 1.3% at furniture and home furnishings stores, 0.6% at electronics and appliance stores, 0.4% at food and beverage stores, and 0.7% at health and personal care stores. Nonstore retailers gained 0.3%, while food services and drinking places rose 1.1%. On the other side, building material and garden equipment and supplies dealers fell 0.6%; sporting goods, hobby, musical instrument and book stores fell 0.5%; general merchandise stores slipped 0.1%; and miscellaneous store retailers declined 3.0%. Clothing and accessories stores were unchanged at the published precision.¹

These are industry estimates, not a statement that every positive or negative number is statistically significant. The Census Bureau explicitly warns that comparisons across industries have not been tested for significance. Ranking furniture stores' 1.3% above electronics stores' 0.6% as a clear business-performance hierarchy would overlook differences in sampling uncertainty. Counting how many industries increased would also overlook their different sizes.¹

Monthly and annual comparisons can point in different directions. Health and personal care sales rose 0.7% from August but remained 1.2% below a year earlier. Sporting goods and the related store categories fell 0.5% during the month but were still up 9.6% year over year. Monthly figures describe the latest step, while annual figures locate sales relative to a year earlier. Establishing whether a rebound becomes a recovery, or a decline persists, requires a sequence of observations.¹

Higher sales dollars do not necessarily mean more purchases

The report measures current-dollar sales. Seasonal adjustment addresses recurring seasonal patterns and calendar differences; price adjustment addresses changes in purchasing power or quantities. They are different operations. Even after seasonal adjustment, higher prices can lift revenue. Conversely, lower prices can allow unit sales to rise while dollar sales change little.¹²

Estimating real retail growth requires price information matched to the sales categories, weights and comparison period, rather than simple subtraction of headline inflation from 0.8%. Food is sold outside food stores, and other products cross retail channels, so store revenue and consumption of a particular commodity do not correspond one to one.²

The coverage has limits too. These are estimates for retail and food services businesses, rather than an account of all service spending. Table 1 specifies that current estimates cover businesses with paid employees only. Before the benchmark report published in April 2025, advance estimates had included nonemployers. Long-term analysis should use a consistent, updated series rather than stitch together numbers from incompatible historical releases.¹

August's upward revision changes September's starting point

The Census Bureau revised July-to-August sales growth from 1.1% to 1.3%, with the associated sampling margin changing from ±0.4 to ±0.1 percentage points. September's 0.8% increase is calculated from that updated August level. The current picture is continued growth at a slower monthly rate than revised August, rather than falling sales. Establishing a persistent slowdown requires subsequent months.¹

A somewhat wider window shows July–September sales running 5.5% above the same three months of 2025, with a sampling margin of ±0.5 percentage points. That three-month annual comparison and September's 6.3% annual increase cover different observation windows and should be read separately. Both measure nominal sales, rather than price-adjusted consumption growth.¹

The release also notes that the monthly series was revised on September 28 using historical corrections and the results of the 2023 and 2024 Annual Integrated Economic Survey. That annual benchmark update and this release's revision to August are distinct layers of adjustment. For budgets and trend analysis, users should record the data vintage and revisit comparisons when an updated series arrives, rather than treat numbers in an earlier announcement as permanent.¹

The 0.8% signal is clearer than some smaller movements

Using the published margin, the approximate 90% confidence interval for September's overall monthly change runs from a 0.4% increase to a 1.2% increase and excludes zero. By contrast, the release's opening chart marks the 0.2% increase excluding autos and gasoline stations with an asterisk. Its 90% interval includes zero, meaning the sampling evidence is insufficient to conclude that the underlying population change differs from zero. Here, the underlying change is the statistical quantity being estimated; it does not mean inflation-adjusted growth.¹

Both conclusions can hold simultaneously: the evidence that total dollar sales rose is relatively clear, while the smaller increase in the narrower aggregate does not meet the same statistical standard. An interval including zero neither proves zero growth nor establishes a decline.

Sampling error is not the only source of uncertainty. The report also identifies nonresponse, incomplete business coverage, and recording, coding and processing problems. These are not all captured by the ±0.4-point headline margin. Nor does that margin describe the range within which any individual company's sales must move.¹

Why advance estimates change, and how businesses can use them

The advance survey draws about 4,800 employer businesses in retail and food services from the larger monthly survey. Within each detailed industry, statisticians calculate a ratio using weighted sales from businesses with usable responses for both the current and previous months. They link that ratio to the previous month's preliminary estimate from the larger sample, then aggregate detailed industries into broader groups. Most nonrespondents are not individually imputed in the advance survey; a limited number of influential companies may receive estimates based on historical performance.¹

This design delivers information early and helps explain why subsequent information can alter the results. Businesses using the figures for purchasing decisions should cross-check their own orders, customer traffic, selling prices and inventory turnover. If rising revenue mainly reflects higher unit prices with unchanged volumes, increasing physical inventory in line with nominal sales could produce excess stock. If quantities keep rising across several channels, the case for replenishment becomes stronger. This is conditional business analysis, not a finding in the report that inventories have accumulated.

Restaurant profit trends also require cost information; household purchasing power and financing conditions require income and credit evidence respectively. Combining these with national industry figures and a business's own operating records helps assess whether sales growth is sustainable.

The next report must test whether growth spreads

According to the release, October's advance estimates are scheduled for November 17 at 8:30 a.m. Eastern Standard Time. Useful checks will include whether September's industry figures are revised, whether growth outside gasoline stations persists, whether the increase excluding autos and gasoline gains clearer statistical support, and whether matched price and quantity evidence points to greater purchasing activity.¹

September has already supplied a clear but bounded conclusion: nominal U.S. retail and food services sales continued to rise, and gasoline stations accounted for most of the month's dollar increase. Broader demand resilience needs to be tested against changes excluding particular industries, subsequent revisions and appropriately matched price information.

Source notes

1. October 15, 2026, 8:30 a.m. EDT · Advance Monthly Sales for Retail and Food Services, September 2026, CB26-165, including tables 1–3 and technical notes · U.S. Census Bureau. This is the current-report endpoint; the cited edition was released October 15, 2026 · Preserved original September 2026 report PDF. ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷ ↩⁸ ↩⁹ ↩¹⁰ ↩¹¹ ↩¹² ↩¹³ ↩¹⁴ ↩¹⁵ ↩¹⁶

2. Accessed October 17, 2026 · Monthly Retail Trade — Definitions, entries on sales, dollar values and kind of business · U.S. Census Bureau. ↩ ↩² ↩³