Construction employment rose in 177 of 360 metro areas and metro divisions between August 2025 and August 2026, fell in 126 and held flat in 57, according to a ranking the Associated General Contractors of America (AGC) published on September 30. AGC built the ranking from the Bureau of Labor Statistics metro employment data released the same morning. Those figures are preliminary and not seasonally adjusted, so AGC compares August with August. Nationally, BLS counts 8,359,000 construction jobs in August 2026, seasonally adjusted and preliminary. That is 22,000 more than July and 120,000 more than August 2025, a 1.5% gain. It is the highest count in the 25 months shown below and sits about 58,000 above the 12-month average of 8,301,500.
National construction employment reached its highest count in 25 months
All employees, construction, thousands, seasonally adjusted
| Month | Employment (000s) | MoM change (000s) | YoY |
|---|---|---|---|
| Aug 2025 | 8,239 | -22 | 0.0% |
| Sep 2025 | 8,255 | 16 | 0.0% |
| 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 | 8,359 | 22 | 1.5% |
What moved
Just under half of the metros shared in the national gain. Houston-Pasadena-The Woodlands added the most construction jobs, 14,100, for a 6% increase to 269,000. Baton Rouge added 13,300 and had the largest percentage gain of any metro, 29%, which took it from 45,800 to 59,100 jobs. St. Louis (11,200 jobs, 14%) and Columbus, Ohio (10,100 jobs, 16%) came next. AGC also lists Davenport-Moline-Rock Island at 14% and Shreveport-Bossier City at 12% among the largest percentage gains, both on much smaller bases.
The largest loss was in the Atlanta-Sandy Springs-Roswell metro division, down 4,900 jobs or 4%. Riverside-San Bernardino-Ontario lost 4,600, the Jersey City-White Plains area lost 4,300, and the Oakland-Fremont-Berkeley division lost 4,000. Pittsburgh and the Lakewood-New Brunswick division in New Jersey each lost 3,500, a 6% drop.
| Metro area or division | BLS series | Aug 2025 | Aug 2026 | Change | Change (%) |
|---|---|---|---|---|---|
| Houston-Pasadena-The Woodlands TX | Construction | 254,900 | 269,000 | 14,100 | 6% |
| Baton Rouge LA | Construction | 45,800 | 59,100 | 13,300 | 29% |
| St. Louis MO-IL | Mining, logging and construction | 81,300 | 92,500 | 11,200 | 14% |
| Columbus OH | Mining, logging and construction | 62,700 | 72,800 | 10,100 | 16% |
| Minneapolis-St. Paul-Bloomington MN-WI | Mining, logging and construction | 101,500 | 108,300 | 6,800 | 7% |
| Phoenix-Mesa-Chandler AZ | Construction | 178,800 | 185,600 | 6,800 | 4% |
| Pittsburgh PA | Construction | 61,800 | 58,300 | -3,500 | -6% |
| Lakewood-New Brunswick NJ (div.) | Mining, logging and construction | 55,800 | 52,300 | -3,500 | -6% |
| Oakland-Fremont-Berkeley CA (div.) | Construction | 72,800 | 68,800 | -4,000 | -5% |
| Jersey City-White Plains NY-NJ | Mining, logging and construction | 72,100 | 67,800 | -4,300 | -6% |
| Riverside-San Bernardino-Ontario CA | Construction | 111,800 | 107,200 | -4,600 | -4% |
| Atlanta-Sandy Springs-Roswell GA (div.) | Construction | 109,900 | 105,000 | -4,900 | -4% |
AGC ties the split to project type. Chief economist Ken Simonson said, “Data centers, power projects and advanced manufacturing plants are spurring employment gains in many metros.” He attributed losses elsewhere to tariffs, materials costs and immigration policy discouraging owners. The AGC release does not say which projects drove the gain in any single metro, so the link between a given metro and a given project type is AGC’s general reading.
The BLS release puts some of these numbers in scale. Houston’s total nonfarm payrolls grew by 44,700 over the year, the most of any metro, so construction accounted for roughly 32% of that gain. Portland-Vancouver-Hillsboro lost 18,800 nonfarm jobs, and 2,600 of them were construction jobs.
How to use this number
Metro construction employment is the closest public count of the labor pool your crews and subcontractors draw from. Two uses:
- Compare your own field headcount change from August 2025 to August 2026 with your metro’s change. If your metro grew 6% and your self-performed headcount is flat, you are hiring against local competitors who have been adding crews all year. If the metro lost jobs and you held headcount, you have gained share of a smaller local workforce, and hiring may come easier than it did a year ago.
- Before you price labor-heavy work in a metro you do not usually work in, look up that metro in AGC’s full list. A 29% jump like Baton Rouge’s means contractors there added more than a quarter to their workforce in one year. Build that into your wage assumptions and into the number of subs you expect to bid.
AGC posts the full table alphabetically and by rank, so you can find your metro. Use the year-over-year change, because the metro figures are not seasonally adjusted and a July-to-August change would mix in normal seasonal hiring.
Revisions and caveats
The August metro figures are preliminary. The BLS technical note describes the current month as a preliminary estimate and the month before it as final. BLS also benchmarks all the estimates once a year to a full count of jobs taken from unemployment insurance tax records, and that benchmark can change a metro’s level for past months.
Three details in AGC’s method matter for reading the ranking:
- In most metros BLS publishes construction only as part of a combined mining, logging and construction series, because samples are small. AGC counts all of that employment as construction. St. Louis, Columbus and Jersey City are in that group, so part of their totals may be mining or logging jobs.
- BLS rounds metro employment to the nearest 100. AGC therefore ranks percentage changes to the nearest whole percent, and many metros tie. Walla Walla, Wash. has the steepest percentage loss at 8%, which is a change of 100 jobs.
- The ranking counts metro divisions separately and leaves out the parent metro totals, to avoid double counting. The Atlanta figure covers the Atlanta-Sandy Springs-Roswell division.
BLS tests over-the-year changes for statistical significance at the 90% confidence level only for total nonfarm payrolls in this release. It publishes sampling error for more detailed series such as construction only on request. So a change of a few hundred jobs in a small metro could be within the survey’s normal error (the release lists 373 of 387 metros as essentially unchanged in total nonfarm jobs). The national figure in the chart is seasonally adjusted and comes from a different series, so compare it with the metro figures only on a year-over-year basis.