Productivity in the construction industry has a direct impact on housing costs and overall economic growth. Understanding its determinants is therefore crucial.
In Barcelona School of Economics Working Paper 1467, “Why Has Construction Productivity Stagnated? The Role of Land-use Regulation,” Leonardo D’Amico, Edward L. Glaeser, Joseph Gyourko, William R. Kerr and Giacomo A. M. Ponzetto study the evolution of construction productivity in the United States over the 20th century.
The authors’ analysis focuses on the relationship between increasingly restrictive land-use regulations and declining house-building productivity since the 1970s. Tighter regulation limits the size of construction projects, which in turn reduces the size of construction firms. Unable to reap economies of scale, developers invest less in innovation. As a consequence, productivity stagnates.
The Puzzle of Declining Construction Productivity
Productivity in the US residential construction sector has declined over the last five decades and is now as low as it was before World War II. This pattern is especially striking because over the same period, productivity rose rapidly in other sectors of the US economy.
In their paper, the authors document a Kuznets-curve pattern (Figure 1). Construction had mirrored the general post-war increase in productivity until the early 1970s, but it uniquely started declining afterwards. What factors drove this reversal?
Note. The figure plots housing units started annually per employee in the construction sector, from 1900 to 2023. We take housing units between 1900 and 1959 from the Macrohistory Database, specifically from the sources denominated US Number of New Private Nonfarm Housing Units Started, One-, Two-, and Three-or-more. From 1959 onward, the data are from the Census’s New Residential Construction program, specifically from the New Privately Owned Housing Units Started series. Employment data in the construction sector between 1900 and 1945 are from the Historical Statistics of the United States, 1789–1945, series D62-76. For the 1929–1945 time period, we also consulted a Bureau of Labor Statistics (BLS) historical report, which corroborates our main series. From 1939 onward, employment in the construction industry is from the BLS’s Current Employment Statistics (CES).
The Impact of Land-Use Regulation
Productivity in construction and in manufacturing sectors such as the auto industry started to diverge in the early 1970s. This timing coincides with the rise of restrictive land-use regulations, which took off after 1973 (Figure 2 below). The authors’ analysis shows how such regulation can have negative unintended consequences for home-building productivity.
Note. The red line (bottom in 2010) plots the log of the ratio between the index of housing units per employee and the index of cars per employee (reported separately in Figure 2). The dark yellow line plots the number of land-use cases per capita, an index of land-use regulation from Ganong and Shoag (2017). Cross-shaped markers are used to denote years in which the denominator in the housing units per employee series was estimated through an out-of-sample forecast (see Appendix C.3 of the Working Paper for details).
Land-use controls restrict the ability to build on a large scale, favoring instead artificially small projects. Developers then face a span-of-control problem: they could scale up efficiently by managing few large projects, but supervising many small ones is costlier.
As a consequence, restrictive project regulation stunts construction firms: they become smaller, less profitable, and thus fewer. The aggregate amount of homebuilding declines and house prices rise. Homebuilders’ economies of scale are reduced, and their incentives for technology investment are blunted.
Evidence from Project- and Firm-Level Data
This study explores empirically the connection between land-use regulation and home-building productivity using detailed data on construction projects, firms and regulatory constraints. A wide range of facts bear out the authors’ theoretical predictions.
Most development projects are very small, and the share of homes built as part of large projects has declined over time. Construction firms are exceptionally small relative to other goods-producing firms, and smaller builders are less productive. The rare developers with more than 500 employees produce more than four times as many houses per employee as the smallest firms.
Areas with stricter land use regulation have particularly small and unproductive construction establishments. A one-standard-deviation increase in the Wharton Residential Land Use Regulatory Index is associated with a one-third reduction in the share of employment in large firms.
Data on patent activity are also consistent with the idea that the rise of land-use regulation contributed to the observed decline in construction productivity. Patenting in construction, and in manufacturing industries that are important suppliers to it, stagnated and diverged from other sectors since the 1970s.
Policy Reform Could Lead to Significant Productivity Gains
To conclude, this BSE Working Paper links the post-1970 decline of US construction productivity to the simultaneous rise of land-use controls. By limiting project size and hindering the growth of construction firms, these regulations have reduced economies of scale and stifled innovation.
The authors’ findings underscore the potential for significant productivity gains in the construction sector through policy reform. A rough calculation shows that if only half of the link between firm size and productivity is causal, US homebuilders would be 60% more productive if their size distribution matched that of manufacturing.
