Coverage: economic releases and developments on Tuesday, November 17, 2026, New York time. The retail and industrial statistics cover October; the GDPNow figure is an estimate for the fourth quarter, not a published GDP result.
Retail spending rises modestly, with statistical uncertainty
The [Census Bureau](https://www.census.gov/retail/marts/www/marts_current.pdf?release=20261117) estimated October retail and food-services sales at $748.2 billion, up 0.2% from September and 6.8% from a year earlier. The monthly estimate carried a plus-or-minus 0.4 percentage-point margin at the stated confidence level. September sales were revised to $746.6 billion. These are seasonally adjusted dollar values, and the report explicitly does not adjust them for price changes.
The uncertainty is part of the result, not a footnote to discard. Because the monthly confidence interval includes zero, the report does not establish a statistically significant increase at that confidence level. That does not prove sales were unchanged; it means the survey cannot distinguish the small estimated movement from zero with the stated degree of confidence. A sensible reading keeps the estimate and its uncertainty together rather than treating the decimal point as perfect precision.
What a sales increase can—and cannot—tell households
The Census tables showed a stronger 0.6% gain excluding motor vehicles and 0.5% excluding both vehicles and gasoline. Those comparisons help reveal the role of major categories, but removing categories changes the question being answered. They are not competing estimates of exactly the same basket. The report also cautions that comparisons across industries have not been tested for statistical significance.
For an illustrative household buying the same groceries at higher prices, dollar spending increases without a larger basket of food. Another household might trade down to cheaper products and purchase more units with the same budget. Retail totals combine many such decisions. To assess living standards or real consumption, readers need price adjustment and broader spending evidence. The retail release is valuable, but it cannot by itself show how purchasing power changed for a typical family.
This also matters when interpreting a retailer’s results. Company sales may change because of new locations, acquisitions, product mix or customer traffic, while the national survey measures a broader universe. A company gaining share can grow faster than its market, and a business losing share can struggle during an expansion. Neither case makes the national data wrong; the two measurements simply operate at different levels.
For small businesses, the practical implication is to compare like with like. A monthly revenue gain should be assessed alongside units sold, selling prices and costs before drawing a conclusion about profitability. More revenue can coexist with a thinner margin. That is an accounting relationship and an analytical question for individual firms, not a claim that Tuesday’s survey directly measured every business’s financial health.
Industrial weakness was not uniform across factories
The [Federal Reserve’s October release](https://www.federalreserve.gov/releases/g17/20261117/g17.pdf) reported a 0.4% decline in total industrial production. Utilities dropped 4.3% and mining fell 0.8%, while manufacturing increased 0.3%. Within factories, motor vehicles and parts rose 4.4%; manufacturing excluding that category edged down 0.1%. The composition matters because the total industrial index includes manufacturing, mining, and electric and gas utilities.
A lower aggregate index therefore does not automatically mean every factory is cutting output. Nor does a positive manufacturing total demonstrate broad strength across all industries. A concentrated increase can lift a composite measure even when other parts are flat or weaker. Readers examining employment or supplier demand should look for the relevant industry rather than assuming the national total describes every employer’s order book.
Capacity utilization stood at 76.4%, three percentage points below its long-run average. That is a measure of production relative to estimated capacity, not the share of workers who are employed. The Fed also scheduled an annual revision for November 24. The current release should therefore be read as the best published estimate at this point, with future revisions capable of changing the historical comparison.
In economic terms, unused capacity and inflation can coexist when pressures differ across sectors or inputs. A factory with spare machinery may still face expensive components, financing or transportation. This is why a single utilization reading cannot resolve the inflation debate. It provides one part of the supply picture, while household prices and business costs require their own evidence.
A growth estimate is not the final score
The [Atlanta Fed’s GDPNow update](https://www.atlantafed.org/en/research-and-data/data/gdpnow/current-and-past-gdpnow-commentaries) put estimated fourth-quarter real GDP growth at a 3.4% seasonally adjusted annual rate on November 17, down from 3.5% on November 9. This is a model estimate for an unfinished quarter. It is neither an official GDP release nor a commitment that growth will eventually be reported at that number.
Annualization also needs care: a 3.4% annualized quarterly pace does not mean output rose 3.4% in a single quarter. The conversion expresses a quarterly pace as if it continued for a year. As more source data arrive, a nowcast can move even without any revision to previously released figures. Its utility is in organizing incoming evidence, while its limitation is that some of the quarter has not yet been observed.
Company reports add detail without replacing the macro data
[Home Depot’s Tuesday release](https://ir.homedepot.com/news-releases/2026/11-17-2026-110221331) reported quarterly sales growth of 4.9% and comparable-sales growth of 2.2%. These are company results over a fiscal quarter, not an alternative estimate of October national retail spending. They add a home-improvement perspective to the broader consumer picture, while differences in coverage and time period prevent a direct one-for-one comparison.
The next useful step is to examine how revenue converts into earnings and cash, and whether demand is broadening across products and customers. At the policy level, [Musalem’s Tuesday remarks](https://www.stlouisfed.org/from-the-president/remarks/2026/remarks-on-us-economy-and-monetary-policy-argentina) emphasized inflation risks rather than declaring the growth data sufficient grounds for a rate decision. The combined evidence calls for a differentiated reading: modest consumer spending growth, uneven industrial activity and continuing uncertainty over the interest-rate path.
