The Industry Needs 349,000 Workers. It Paid the Ones It Has Less Than Last Year.

In January, Associated Builders and Contractors published its annual workforce projection: the construction industry needs 349,000 net new workers in 2026, a number its chief economist Anirban Basu called "sizable" and one that immediately landed in trade-publication headlines, conference keynotes, and vendor pitch decks. Nobody on any of those stages mentioned what the Bureau of Labor Statistics was simultaneously reporting about wages.

Real wages for residential building workers fell 1.2 percent year-over-year through March 2026, according to BLS data analyzed by the National Association of Home Builders in May. Nominal earnings grew just 2.1 percent, a collapse from the 9.4 percent peak reached in mid-2024 that actually looked like what economists expect when employers are genuinely competing for scarce labor.

Economics has a clean test for whether a shortage is real: wages rise until the market clears, because employers bid against each other and people doing something else show up once the money gets good enough. A 9.4 percent nominal growth rate looks like that test being passed. A 2.1 percent rate with real purchasing power declining does not, and ABC's own trajectory tells the same story, with the projected gap dropping from above 500,000 in 2023 to 439,000 in 2025 to today's 349,000, a 30 percent decline that Basu attributed primarily to retirements and "modest" spending forecasts rather than any influx of new workers.

February 2026 removed remaining ambiguity. Construction's hiring rate fell to 3.3 percent, the lowest since the BLS began the Job Openings and Labor Turnover Survey in December 2000, while layoffs held at 1.8 percent and quits at 1.5 percent, producing what Basu himself called "the month with the least construction labor force churn" in the survey's 25-year history. Residential lost 3,100 jobs in December 2025, open positions have been trending downward, and the share of contractors expecting to expand headcount dropped from 69 to 63 percent while those planning cuts rose from 10 to 15 percent, per AGC and Sage.

Where AI Fits, and Where It Doesn't

When industry groups cite the labor gap, the next sentence almost always involves technology: AI-powered scheduling, computer vision for progress tracking, robotic prefabrication, all pitched as the answer to a workforce that will not materialize. Tools like Buildots and OpenSpace genuinely reduce the person-hours required for site documentation, which means a smaller crew can cover more ground in the same number of days.

But if AI-driven productivity gains were truly filling the gap, you would expect rising output per worker alongside rising wages, because workers in a tight market capture some share of the surplus their higher productivity generates. Output holding steady while real wages decline means the gains are flowing to builder margins, not to the crews pouring foundations and hanging drywall, which is a coherent business strategy but not a solution to a labor shortage in any sense a worker would recognize.

And the retirement problem Basu identified as the primary demand driver sits outside what current AI tools address. When a plumber with 30 years of diagnostic intuition retires, no scheduling algorithm absorbs the feel for whether a joint is good, the ear for a compressor running wrong, the instinct that something behind the drywall is off.

What Aggregate Numbers Hide

Licensed electricians, experienced plumbers, certified welders, and lead framers capable of training apprentices are measurably harder to recruit than general laborers, and their wages may be climbing even as the residential composite falls, because aggregate averages flatten genuine scarcity into a single misleading figure. Regional variation compounds the distortion: union markets in Chicago and the Pacific Northwest look nothing like nonunion rates across the South. Nobody hires nationally; they hire from the trucks that show up Monday morning.

For builders planning 2027 headcount, the actionable signal is not the 349,000 headline but the wage trajectory underneath it: experienced tradespeople are aging out, no vendor product replaces that loss, and retention of existing workers matters more than recruitment when purchasing power is eroding. If you are buying or building a home, understand the framing, because when your builder says "we can't find anyone," the BLS data suggests the more honest version may be "we can't find anyone at the price we want to pay."

Limitations

NAHB's wage analysis covers NAICS 2361 (residential building construction, including new single-family and remodelers) but excludes specialty trade contractors under NAICS 238x, whose wage patterns may diverge. ABC's 349,000 is a model-based projection, not a census. December employment losses carry seasonal effects, and no BLS dataset captures informal or undocumented labor, which affects supply-side dynamics in ways that remain invisible in every number cited above.