American manufacturing productivity has tripled since 1948. Construction productivity tells the opposite story. Brian Potter's Construction Physics analysis shows U.S. construction labor productivity is essentially unchanged from 1948—and has declined more than 30 percent since its 1970 peak. This isn't measurement error: physical measures like housing units per worker show the same stagnation. While manufacturing saw sustained gains, construction moved sideways, then backward.
The productivity collapse translates directly into stubborn costs. Potter's examination of construction cost indexes shows costs rising at or above overall inflation across decades. Unlike manufacturing, where productivity gains drive prices down even when input costs rise, construction offers no such relief. When steel or lumber prices fall, construction costs rarely follow. The industry has lost the ability to convert input savings into output affordability.
Planning timelines have stretched as construction has slowed. Potter's analysis of 67 major U.S. bridges since 1900 reveals that bridge planning alone now averages ten years before construction begins—compared to six years in the 1960s and 1970s. The Golden Gate Bridge was constructed in four years starting 1933, though planning had begun in 1921. Construction times themselves have worsened: bridges built between 1980 and 1999 took 6.5 years to construct on average, more than double the 3.1 years typical of 1940–1959 bridges.
The absence of economies of scale deepens the puzzle. Despite industry consolidation—Lennar and D.R. Horton each now build over 80,000 homes annually—larger builders demonstrate no cost advantage over small operators. A Harvard Joint Center for Housing Studies analysis found that large homebuilders actually had higher construction costs than smaller competitors in the late 1990s and early 2000s. Recent data confirms this persists: builders producing 1,000 homes annually show similar gross margins and selling prices to those building 80,000. The U.S. has over 65,000 homebuilding firms, with the four largest holding just 18 percent market share—compared to 90 percent concentration in aircraft manufacturing.
The fragmented structure reflects genuine constraints, not inefficiency. In U.S. homebuilding, the difference between raw input costs and final selling prices is already small, leaving little margin for scale economies to close. Fixed costs can't be spread dramatically thinner when material and labor already dominate expenses. Learning-by-doing effects remain modest. The result: no manufacturing-style cost curve where doubling production cuts unit costs significantly. Homebuilding resists the fundamental mechanisms that drive progress elsewhere.
Regulation emerges as the primary explanation. An NBER working paper by D'Amico, Glaeser, Gyourko, Kerr, and Ponzetto documents that local land-use and federal environmental regulations proliferated in the early 1970s—precisely when construction productivity began its sustained decline. Their analysis of the Wharton Residential Land Use Regulatory Index shows that metropolitan areas with stricter regulations have smaller, less productive construction firms and reduced building activity. Goldman Sachs Research estimates that a 1 percent increase in regulation intensity lowers annual construction productivity growth by 0.6 percentage points. The cumulative effect: increased U.S. regulation since the 1960s has likely reduced annual construction productivity growth by 0.7 percentage points, offsetting nearly all the estimated 0.8 percentage point boost from technological change.
Regulatory burdens favor small projects, artificially capping firm size. Potter's analysis using CoreLogic microdata covering 167 counties shows that in 1950–51, 37 percent of homes were built in the largest 1 percent of development projects; by 2018–19, only 24 percent. As regulatory burdens grow, approved projects shrink and profitability declines. Building in highly regulated areas becomes unattractive, reducing developer entry and competition. The NBER researchers calculate that if half the observed link between productivity and firm size is causal, homebuilding would be approximately 60 percent more productive if U.S. firm size distribution matched manufacturing.
Innovation has withered alongside scale. The share of industrial machines in total construction costs spiked from 4 percent to 12 percent between 1948 and 1968, then fell to 10 percent by the 1970s and stayed there. Prefabrication technology accounted for one-third of new residential units at its 1960s peak; it now represents just 5 percent. Patent activity in construction and construction-supply sectors has lagged the broader economy, especially since the 1970s. Major technological breakthroughs of recent decades—concentrated in information, communication, and chemical manufacturing—are distant from construction's supply chain, yielding minimal spillover benefits.
Construction alone accounted for roughly one-third of the decline in U.S. trend GDP growth since World War II—a loss equivalent to approximately $1 trillion every five years.
The macroeconomic consequences are severe. A 2022 study concluded that construction alone accounted for roughly one-third of the decline in U.S. trend GDP growth since World War II—a loss equivalent to approximately $1 trillion every five years. A 2023 working paper estimates that if construction productivity had grown at just 1 percent annually since 1970, aggregate labor productivity growth would be 0.18 percentage points higher, leaving current productivity and income per capita roughly 10 percent above actual levels. Construction represented 4.5 percent of GDP in 2024, but its productivity drag punches far above that weight because construction supplies capital to every other sector.
Casey Handmer's analysis suggests this is a developed-economy phenomenon, not uniquely American. Examining Australia, he documents zero public sector productivity improvement since 2001, with construction stagnation compounding other economic challenges. Handmer argues the productivity crisis risks trapping Australia in increasingly interventionist policy cycles by 2040: governments subsidize unaffordable housing, then impose price controls when subsidies fail, then consider nationalization when controls create shortages—a pattern he sees emerging in healthcare and education. The pattern: regulation restricts supply, prices rise, political pressure mounts, further intervention compounds problems.
Build-to-rent housing has emerged as an adaptive response within existing constraints. Potter documents that BTR rose from under 2 percent of housing starts in the 1990s to over 7 percent today—at least 68,000 units in 2025, possibly exceeding 100,000. The modern BTR industry arose from the 2008 financial crisis, when firms like Blackstone's Invitation Homes and American Homes 4 Rent bought foreclosed properties at scale to rent to Americans shut out of ownership by tightened lending standards. By 2022 the four largest rental companies owned over 200,000 homes. BTR represents a shift in ownership structure that works around—but doesn't solve—construction productivity problems.
The timing creates significant challenges. Electrification, data centers, and defense manufacturing all require rapid physical buildout. The energy transition depends on constructing generation capacity, transmission lines, and charging infrastructure at unprecedented speed. AI data centers need massive new facilities; reports indicate companies like Meta have explored accelerated construction approaches to speed deployment. Reshoring semiconductor fabrication demands new fabs. Every industrial policy ambition confronts the same constraint: a half-century of compounding productivity decline has created a physical bottleneck that capital and political will cannot easily bypass.
The core problem remains unsolved because solutions require confronting entrenched interests. Reducing regulatory burdens means overriding local control and environmental review processes with powerful constituencies. Enabling larger projects and firms means accepting concentration that triggers political opposition. Permitting innovation means accepting risks that current processes are designed to eliminate. The productivity collapse is not a mystery—its causes are documented and understood. But understanding doesn't translate to action when action requires dismantling five decades of accumulated regulatory infrastructure. Construction productivity is a political problem disguised as an economic one, and politics has so far chosen stagnation.