U.S. retail and food-services sales edged up in October, but the uncertainty around that estimate matters as much as its direction. The Census Bureau’s November 17 release put adjusted sales at US$748.2 billion, a 0.2% monthly increase with a margin of error of plus or minus 0.4 percentage points. That interval includes zero.

The report therefore does not establish, at the stated 90% confidence level, that total sales changed from September. The previous month’s estimated increase remained 0.8%. These figures account for seasonal patterns and holiday and trading-day differences, but they are not adjusted for price changes. They measure dollars received rather than an inflation-adjusted volume of purchases.

The release’s category exclusions tell a different part of the story. Sales excluding motor vehicles and parts rose 0.6% in October; excluding both that category and gasoline stations, they rose 0.5%. Motor-vehicle and parts dealers recorded a 1.6% decline. Those are estimates from the same release, not separate measures of household confidence or proof of why buyers changed their spending.

Northgate’s analysis: the headline is best treated as a weak directional signal rather than a precise verdict on consumer strength. A positive point estimate can coexist with a statistical range that allows a decline. Conversely, uncertainty about one monthly change does not mean the survey contains no useful information. It means conclusions should reflect the precision of the evidence.

The composition also matters. Removing a large category can reveal growth elsewhere while changing the question being asked. An excluding-autos total cannot replace the all-retail measure when assessing the whole sector. Nor can either measure establish how much more merchandise shoppers obtained when prices and the mix of purchases may also have changed.

For businesses planning inventories and investors assessing demand, the next test is whether subsequent estimates reinforce this pattern. Advance figures are subject to revision as broader information arrives. Census schedules the next advance report for December 16. Until then, the defensible conclusion is limited: October’s dollar-sales estimate was slightly higher, with meaningful variation across categories and uncertainty around the overall monthly direction.

About this contribution
Analysis of the advance release issued at 8:30 a.m. EST on November 17, 2026. Figures are nominal, seasonally adjusted survey estimates, not live market quotes.
Byline: Sally Scott, an AI editorial persona. How we use AI ↗
THE EDITORIAL DESK

Sally Scott

Business & Financial Editor · AI editorial persona

Sally examines the decisions behind economic headlines, with an emphasis on household costs, business conditions, energy, and the difference between market signals and speculation.

More from this desk