← Back to Library

Trends in US construction productivity

Brian Potter · Construction Physics · ·31 min read
Commentary by Hex Index staff

The Sixty-Year Stagnation

For decades, American construction has fallen behind almost every other major sector in productivity growth. Brian Potter's exhaustive analysis pulls together six decades of data to show something remarkable: a worker in 2020 produced less than a worker in 1970. The numbers don't just flatline — they deteriorate.

Sector-Wide Decline

Brian Potter writes, "construction holds the dubious honour of having the lowest productivity gains of any industry." The Economist's assessment from 2017 captures the breadth of the failure. Paul Teicholz, a Stanford civil engineering professor, calculated that between 1964 and 2004, construction productivity declined by 0.59% per year on average. This was "particularly alarming when compared to the increasing labor productivity in all non-farm industries, which have experienced an increasing productivity of 1.77% per year over the same time period."

Trends in US construction productivity

The gap widens when economists Austan Goolsbee and Chad Syverson measure total factor productivity — accounting for both labor and capital inputs. Brian Potter notes their finding: "Productivity increased from the 1950s until the mid-1960s, but since then it has declined by roughly 50%." That 2025 paper has become the reference point for Federal Reserve briefs, Goldman Sachs reports, and Ezra Klein's 2023 New York Times column on the subject.

A construction worker in 2020 produced less than a construction worker in 1970, at least according to the official statistics.

Critics might note that sector-wide metrics conflate vastly different types of construction — hospitals, highways, single-family homes — into one aggregate. The Bureau of Labor Statistics data shows more nuance: multifamily and industrial building productivity actually rose, while single-family and highway work stagnated or declined.

The Housing Puzzle

Subsector metrics reveal contradictions. Brian Potter points out that "single-family construction shows an increase in productivity of close to 50% between 1990 and 2020" when measured by square feet per employee. Yet the same subsector shows unchanging productivity when measured by spending per labor hour. For multifamily construction, the reverse holds true.

This inconsistency matters. Princeton economist Leonardo D'Amico and coauthors found housing productivity declined significantly since the 1960s using units-per-employee. But Goolsbee and Syverson's square-footage metric tells a different story. Brian Potter's analysis suggests the measurement itself determines the conclusion.

Task-Level Reality

The most granular data comes from estimating guides — RS Means and Craftsman have published labor requirements since the 1950s. Brian Potter compares Craftsman's 1993 and 2026 estimates for building a typical single-family home: "the estimated number of hours required to build a typical single-family home... are almost identical." One task improved: insulation installation, which took six days in 1993 and now takes three.

Task-level studies show modest gains. Goodrum's 2002 analysis of 200 construction tasks found productivity improved for 107 tasks, unchanged for 64, and declined for 30. Brian Potter's own RS Means comparison from 1985 to 2023 found 10 tasks more productive, 10 less productive, and 20 unchanged.

The companion deep dive on workforce productivity explores why individual task improvements don't aggregate into sector-wide gains. Coordination costs, regulatory layers, and fragmented project management may erase micro-level efficiencies. The BIM Collaboration Format deep dive examines whether digital coordination tools have actually reduced rework and delays — or just added another layer of complexity.

What the Numbers Hide

Brian Potter acknowledges measurement challenges throughout. Inflation adjustments matter — spending data uses the Consumer Price Index, but construction costs may rise differently. Quality adjustments matter too: are modern buildings more complex, with higher safety standards, energy requirements, and accessibility features that demand more labor?

Garcia and Molloy's 2025 paper attempts quality adjustments in their productivity metric. Brian Potter notes this effort but leaves the question unresolved: if we're building safer, more energy-efficient, more accessible structures, does "less output per hour" actually mean less productivity, or just different output?

Bottom Line

Sixty years of data converge on one verdict: American construction productivity has stagnated or declined, with occasional subsector exceptions that don't offset the broader trend. Brian Potter's meticulous aggregation of metrics — from sector-wide to task-level — shows this isn't a measurement artifact but a structural reality. The insulation installation improvement stands as the lone clear gain. For readers who depend on construction costs falling, the data offers no comfort.

Sources

Trends in US construction productivity

by Brian Potter · Construction Physics · Read full article

(This is a chapter of a longer report I’m working on that summarizes and expands the last several years of my work on construction productivity. I plan on publishing one chapter a month on the newsletter, and aim to have the full report done by the end of the year.)

For decades, American construction has fallen behind almost every other major sector in productivity growth. As far back as 1970 researchers noted that construction productivity improvement significantly lagged productivity improvement in the economy overall, and by 1985 economists were investigating what appeared to be declining construction productivity. Stanford civil engineering professor Paul Teicholz noted in a 2004 article in AECbytes that between 1964 and 2004, construction productivity declined by 0.59% per year on average, which was “particularly alarming when compared to the increasing labor productivity in all non-farm industries, which have experienced an increasing productivity of 1.77%/year over the same time period.” A 2017 article in The Economist noted that “construction holds the dubious honour of having the lowest productivity gains of any industry.” In a 2023 New York Times column, Ezra Klein wrote that “A construction worker in 2020 produced less than a construction worker in 1970, at least according to the official statistics.”

The trend of construction productivity in the United States failing to improve over time is indeed concerning. “Productivity” means some measure of output, divided by some measure of input. When productivity is improving, we get more output for a given amount of input over time; if productivity is falling, we get less output for a given amount of input over time. If productivity doesn’t improve, we can’t expect construction costs to fall and things like houses, roads, and bridges to get any cheaper. Because of this, it’s worth looking deeply at what exactly the trends in US construction productivity are.

Economists and researchers measure construction productivity in a variety of different ways. We can broadly categorize these metrics by their level of granularity:

At the lowest level of granularity, we have metrics that track productivity changes across the entire construction sector.

Slightly more granular are metrics that look at productivity changes in a particular subsector, such as housing construction.

Looking more specifically, we have metrics that look at productivity changes for constructing particular buildings.

And finally we have metrics that track productivity changes for individual construction tasks.

Each category of metric gives a slightly different ...