Construction employment fell by 3,000 in October 2026 to 8,367,000, according to the Employment Situation report the Bureau of Labor Statistics released on November 6. The figure is seasonally adjusted and preliminary. It is 0.04% below September and 124,000, or 1.5%, above October 2025. October ranks second among the 25 months in the chart, behind the revised September figure, and sits about 46,000 above the 12-month average of 8,321,000. BLS also revised September up to 8,370,000 from the 8,364,000 it first published, so the September gain is now 16,000 instead of 11,000.
Construction employment held near its 25-month high in October
All employees, construction, thousands, seasonally adjusted
| Month | Employment (000s) | MoM change (000s) | YoY |
|---|---|---|---|
| Oct 2025 | 8,243 | -12 | -0.2% |
| Nov 2025 | 8,279 | 36 | 0.1% |
| Dec 2025 | 8,272 | -7 | -0.1% |
| Jan 2026 | 8,317 | 45 | 0.6% |
| Feb 2026 | 8,296 | -21 | 0.3% |
| Mar 2026 | 8,311 | 15 | 0.5% |
| Apr 2026 | 8,314 | 3 | 0.5% |
| May 2026 | 8,316 | 2 | 0.6% |
| Jun 2026 | 8,319 | 3 | 0.6% |
| Jul 2026 | 8,337 | 18 | 0.9% |
| Aug 2026 | 8,354 | 17 | 1.4% |
| Sep 2026 | 8,370 | 16 | 1.4% |
| Oct 2026 | 8,367 | -3 | 1.5% |
What moved
BLS put construction on its list of industries where employment “showed little change over the month.” The detail in table B-1 shows where the 3,000 came from. Residential specialty trade contractors lost 15,200 jobs, the largest move of any construction line, and residential building construction lost 1,200. The three nonresidential lines added 13,300 between them, with 7,900 at nonresidential specialty trade contractors, 2,900 at nonresidential builders and 2,500 in heavy and civil engineering.
| Segment | Oct 2025 (000s) | Sep 2026 (000s) | Oct 2026 (000s) | Change from Sep (000s) | Change from Oct 2025 (000s) |
|---|---|---|---|---|---|
| Residential building construction | 925.3 | 922.1 | 920.9 | -1.2 | -4.4 |
| Nonresidential building construction | 918.3 | 949.7 | 952.6 | 2.9 | 34.3 |
| Heavy and civil engineering construction | 1,183.5 | 1,210.2 | 1,212.7 | 2.5 | 29.2 |
| Residential specialty trade contractors | 2,368.2 | 2,347.6 | 2,332.4 | -15.2 | -35.8 |
| Nonresidential specialty trade contractors | 2,847.5 | 2,940.7 | 2,948.6 | 7.9 | 101.1 |
The 12-month view runs the same way and by a wider margin. The three nonresidential lines have added 164,600 jobs since October 2025, and the two residential lines have lost 40,200. Nonresidential specialty trade contractors grew 3.6% to 2,948,600 and supplied 101,100 of that gain on their own. Residential specialty trade contractors are down 1.5% to 2,332,400, and October’s 15,200 decline is 0.6% of that segment in a single month. The spending side of the same split is in the September construction spending post.
So, turning to pay, average hourly earnings for all construction employees rose 38 cents in October to $42.03, a 0.9% monthly gain, and are up 4.7% from $40.14 in October 2025. Across all private employers BLS reports a 3.1% rise over the same 12 months. Average weekly earnings in construction reached $1,655.98, up 6.1% from a year earlier, since weekly pay moves with hours worked as well as with the hourly rate.
How to use this number
- Compare your field headcount to your segment. A residential trade contractor whose headcount is down about 1.5% from October 2025 has moved with residential specialty trades as a whole. A nonresidential trade contractor up less than 3.6% has grown more slowly than its segment. Your own payroll count is not seasonally adjusted, so compare October to October, the same 12-month window BLS uses for this figure.
- Check the labor escalation in bids for 2027 work. Construction hourly earnings rose 4.7% over the year, faster than the 3.1% across all private employers. If your estimates carry a lower escalation rate on labor, test it against your own payroll for the same 12 months, and read it next to the September job openings and quits figures. The BLS average covers every trade, region and pay arrangement in construction, so your crews may run above or below it.
Revisions and caveats
October is preliminary, and so is the revised September. BLS will revise both when it publishes November data on December 4, 2026. This month’s revision worked like this:
- On October 2, BLS first published September construction employment at 8,364,000 and August at 8,353,000. FRED’s USCONS series still carried those figures when this post was written.
- The November 6 release revised September to 8,370,000 and August to 8,354,000.
- The September gain is now 16,000, and October’s 3,000 decline is measured from the higher September figure.
If BLS had left September at its first estimate, October’s 8,367,000 would show as a gain of 3,000. The sign of this month’s change comes from the revision. A move of 3,000 in either direction is 0.04% of the total, and BLS grouped construction with the industries that showed little change. BLS says monthly revisions come from reports that businesses and government agencies send in after the first estimate, and from recalculated seasonal factors. Across all industries, BLS revised August and September combined up by 84,000.
Before seasonal adjustment, construction employment rose to 8,525,000 in October from 8,512,000 in September, a gain of 13,000. The seasonally adjusted series shows a decline of 3,000, which means employers added fewer construction jobs in October than the seasonal factors expected. (Unadjusted construction employment was 8,413,000 in October 2025, so the unadjusted count is also up from a year ago.)
The figures come from the Current Employment Statistics survey of employer payrolls. FRED’s series notes say the construction count includes working supervisors, craft workers, apprentices, helpers and laborers, whether they work at the site or in shops and yards on work the construction trades normally perform. The B-1 table notes that the data have been revised to March 2025 benchmark levels, so the history in the chart will move again when BLS applies its next annual benchmark. The chart and the 13-month table come from the BLS data API, because FRED had not yet posted the October figure when this post was written.