Our research investigates the hypothesis that land-use regulation drives the difference between manufacturing and construction productivity partly by ensuring that residential construction is built by smaller and less productive firms. Construction is regulated before building begins, and local laws and state and federal environmental regulations may restrict the ability to begin development at all. The approval process can require years of community outreach, and catering to the wishes of residents is often easier for smaller projects. By contrast, federal manufacturing regulations are typically enforced after production begins, and manufacturers’ large size can make compliance easier. When regulations limit the size of projects, they lead to the creation of smaller firms, which invest less in technology, ultimately slowing productivity growth—as seen in the United States.
Our research finds that productivity growth in home building soared from 1935 to 1970. During this period, the number of homes produced per construction worker often grew faster than the number of cars produced per automobile worker or total manufacturing output per industrial worker. This fact disproves the view that it is impossible to improve productivity in construction—a view that is also at odds with the many innovations introduced by the great post–World War II builders such as William Levitt. However, the construction sector began to stagnate around 1970 and soon lagged the rest of the economy. At that time, the increasingly prosperous United States started embracing land-use controls and other regulations.
The rise of building regulations accounts for at least part of the productivity slowdown in construction. We present this argument through a formal model. In this model, builders choose how much to invest in technology and have limited ability to monitor different projects. We represent regulation as decreasing the probability of obtaining a building permit. The likelihood that a project is approved decreases as its size increases. Thus, regulation causes the average project size to shrink as entrepreneurs pursue a smaller project scale for a higher probability of receiving a permit. The smaller size of projects then leads firms to shrink because entrepreneurs cannot monitor, for instance, 100 projects with two houses each as easily as two projects with 100 houses each. As firm size shrinks, so do incentives to invest in technology. The overall productivity of the industry consequently declines.
Our model explains why regulatory expansion may lead to productivity growth in manufacturing but productivity decline in construction. Regulation of manufacturing typically increases fixed costs and raises barriers to entry. In a competitive market, higher entry costs tend to reduce the number of firms in an industry and increase firm size. Bigger firms have greater incentives to invest in technology, so lower production costs partially offset the adverse effects of restricted entry on consumers. Conversely, the small scale and limited technological investment of the firms that remain in the industry exacerbate the adverse effects of the regulation of construction projects on consumers.
Our research presents evidence that construction projects and firms are indeed quite small. While only 2 percent of manufacturing employees work in firms with fewer than five employees, 40 percent of employment in new single-family housing construction is in such tiny firms. The typical land parcel bought between 2013 and 2018 for single-family development is also quite small. Prior research has found that the median size of 3,600 parcels purchased between 2013 and 2018 across 24 metropolitan areas was 7 acres; 94 percent of parcels had less than 100 acres of land, and virtually no parcels contained more than 1,000 acres.
No dataset existed documenting project size over time, so we created one using an algorithm that estimates development scale by considering homes built near each other in the same two-year period. Our dataset reveals that the largest builders in the 1960s worked with land parcels of more than 5,000 acres and built many thousands of homes on each of them; we confirmed this fact using contemporaneous press reports. Since then, the share of housing built in large projects has fallen by over one-third.
Our research confirms the strong connection between firm size and productivity in construction using data from the Economic Census and the Longitudinal Business Database. These datasets contain direct measures of output, such as homes built per firm. In housing construction, firms with 500 or more employees produced four times as many units per employee than firms with fewer than 20 employees. If half the correlation between size and productivity is causal, then the construction sector would be nearly 60 percent more productive if its distribution of firm sizes were the same as in manufacturing.
Our research also finds that areas with stricter land-use regulations, as measured by the Wharton Residential Land Use Regulatory Index, have smaller and less productive construction firms. The inverse relationship between size and regulation holds across the whole industry, but it is especially strong for firms that construct buildings. For these firms, an increase in the index by one standard deviation—which is approximately the difference between Atlanta’s and San Francisco’s regulatory strictness—is associated with a 12 percent reduction in total revenue per firm and a one-third reduction in the share of employment in large firms. Among all construction firms, increasing the index by one standard deviation is associated with decreasing payroll per employee by 3 percent, capital per employee by 5 percent, and revenue per employee by over 5 percent. Our findings also reveal much less residential and nonresidential construction activity in areas with stricter land-use regulations.
Finally, our research uses data on patents and spending on research and development to test the prediction that smaller firm size is associated with less investment in technology. Patenting levels for construction, manufacturing, and all other industries moved together through the 1950s. After 1970, patents per employee soared in manufacturing but declined in construction. Likewise, corporate research and development spending as a share of revenue is at least 10 times higher in manufacturing than in construction.
Note
This research brief is based on Leonardo D’Amico et al., “Why Has Construction Productivity Stagnated? The Role of Land-Use Regulation,” National Bureau of Economic Research Working Paper no. 33188, November 2024.
