Sixty years of flat to falling productivity.
The rest of the economy roughly tripled output per hour since the mid-1960s. Construction did not move. The finding survives at sector, subsector, building, and task level, and across more than forty countries.
In sixty years of US homebuilding, the one task that got meaningfully faster is installing insulation.
Goolsbee and Syverson (2025) find TFP in construction rises through the mid-1960s and then falls by roughly half.
The finding does not move.
Teicholz measured construction labor productivity from 1964 to 2004 and found it falling at roughly 0.59% per year, while non-farm productivity rose at 1.77%. Goolsbee and Syverson (2025) cross-checked across multiple datasets and arrived at the same shape. The McKinsey Global Institute placed the cumulative multiplier from 1947 to 2010 at 1.1× for construction, against 16.1× for agriculture and 8.6× for manufacturing.
Walk the data down and it never breaks. Sector level, the holdout persists. Subsector level, residential, commercial, and infrastructure each show the same flat line. EU KLEMS data replicates the finding across more than forty countries. And at the task level, Craftsman and RS Means estimating guides show that the labor budget for laying brick, hanging drywall, framing a wall, or pulling wire has barely moved in five decades.
In sixty years of homebuilding, exactly one task got meaningfully faster. Installing insulation. Everything else is roughly the same labor budget as in 1965.
Data synthesized from McKinsey Global Institute, Teicholz, Goolsbee and Syverson (2025), the BLS, the BEA, EU KLEMS, and Craftsman and RS Means estimating guides, plus Aspen Institute and Institute for Progress work by Brian Potter. Series are illustrative and indexed for legibility.