In his New York Times column last week Erza Klein raised the intriguing question as to why productivity in construction is so bad. The piece was appropriately entitled “The Story Construction Tells About America’ Economy is Disturbing.” It is largely based on a recent research paper by economists Austan Goolsbee and Chad Syverson on “The Strange and Awful Path of Productivity in the U.S. Construction Sector.”
Goolsbee and Syverson conclude that the productivity slowdown is real and not, as some would claim, a case of mismeasurement or other problems with the data. As to what is causing the slowdown, they have no solid explanation. As they note, “[M]easurement problems alone likely cannot explain all of the decline and that there are some problems facing productivity in the industry that have not been documented previously.”
Klein offers one possible explanation: increased regulation and other requirements (“paperwork, and paperwork, and more paperwork”). Beyond paperwork, he argues that construction faces obstacles that other industries do not. “When you construct a new building or subway tunnel or highway, you have to navigate neighbors and communities and existing roads and emergency access vehicles and politicians and beloved views of the park and the possibility of earthquakes and on and on.”
Goolsbee and Syverson end with the familiar but in this case completely appropriate recommendation that “Further research is needed to test between competing explanations and sharpen the picture of what has been happening in the sector.”
I have to say that I strongly agree with the need for more analysis. But let me suggest that it might be more illuminating to turn the research question around. Rather than ask why is current productivity low compared with the 50s & 60s, let us ask why was it so high back then? In other words, what were the forces and circumstance that supported this high level of productivity? And can we create new or recreate earlier forces that will have the same effect?
A short vignette from a recent paper by Martin Baily on productivity (based on a McKinsey & Company report Baily worked on) provides some insights. In this case, he was looking at the difference between Brazil and the U.S. in residential construction:
“Productivity was very low in Brazil, only about one-fifth of the U.S. level. The conventional wisdom in Brazil was that this low productivity was the result of the low educational level of the construction workers.
. . .
Instead, the productivity difference arose from two main reasons. First, most U.S. residential construction is carried out in sites where a large area is cleared and then multiple copies of pretty much the same house is built. This allows economies of scale. Second, a U.S. construction site is carefully orchestrated by site managers. Special trade workers, such as plumbers, carpenters and electricians are brought to the site only when required. These workers move from site to site as needed. Utilization of labor is much better in residential construction in the United States.”
In other words, organizational intangible assets were the key in explaining the differences in productivity.
I would argue that these two productivity-enhancing factors of economies of scale and relatively efficient labor utilization may also explain the differences in productivity between then (50s & 60s) and now. Both appeared in the construction industry during the Golden Age. And once they were widely diffused, their ability to raise productivity diminished. So far, nothing major in the way of new organizational intangible assets has come along to boost productivity.
Neither Goolsbee and Syverson nor Klein take these intangible assets into account. Goolsbee and Syverson make an effort, but come up short. They assume the following: “[W]e can use IP capital stock as a proxy for intangible capital in the industry (which would include things like know-how, organizational strength, trade secrets, buyer-supplier relationships, sector-specific human capital, and so on).”
This assumption does not hold for construction. The measure of intangible capital they use is BEA’s data for investment in intellectual property products, which include R&D, software purchases, and artistic originals. As Goolsbee and Syverson note, construction companies have little IP capital using this definition. And, most importantly, had little to no investments in IP capital during the Golden Age of the 50s and 60s. Thus, Goolsbee and Syverson’s analysis shows little impact on productivity due to intangibles, that is ongoing investment in R&D and software. But, as the discussion above indicates, it is likely that construction companies had a high level of organizational (and other) intangible capital at one point, which dissipated over time.
If this is the case, then it is possible that the decline in productivity was due to an inability to maintain the earlier intangible capital or to create a new stock of intangibles. Unfortunately, we don’t have good data and metrics for this broader set of intangibles at the macroeconomic (System of National Accounts – GDP) level. [Which is why I use the same data as Goolsbee and Syverson in my analysis of macro-level investment in what I call knowledge-related areas].
Klein hints at another set of intangibles when he declares that “[T]he frictions are in navigating local regulations, community considerations, neighbors’ qualms and politicians’ interest.” This local knowledge and navigational skills are key intangible assets in and of itself. Based on my own experience as a government official involved in DC’s planning and zoning activities, I can attest to the fact that some companies are better at dealing with this complexity that others. I don’t doubt that the complexity facing construction projects has increased since the 50s and 60s. And I suspect that companies’ ability to navigate this complexity (which is an organizational intangible asset) has not increased accordingly.
Accepting the (unproven) hypothesis that intangibles played a (large?) role in the productivity of the construction industry, we can turn our attention to how to foster intangibles in the industry.
A number of years ago I did a report on the competitiveness of the construction industry for the then Congressional Office of Technology Assessment (part of a larger study on services). Back then, I noted that in contrast to manufacturing, construction was basically a craft-based industry. That study concluded that increase productivity in construction will require companies adopt better techniques on the job site and more extensive use of off-site prefabricated components but that there were major barriers to innovation. The industry suffers from a lack of incentives to adopt new technologies (and other intangibles). For example, as I noted back then, “Not only may superior construction methods be precluded, but the system rewards conservative choices. Once the contractor has won a job with a fixed-price bid, there is little incentive to do anything but follow the specifications the bid was based on.” In addition, companies lack of capability to utilize these intangible due in part the fragmented nature of the industry.
More recently (2017), a report from McKinsey Global Institute confirmed my decades-old findings, including the large potential gains from the use of prefabricated standardized components and the fact that “Most individual players lack both the incentives and the scale to change the system.”
The report lays out the factors behind the productivity problem:
“The industry is extensively regulated, very dependent on public-sector demand, and highly cyclical. Informality and sometimes corruption distort the market. Construction is highly fragmented. Contracts have mismatches in risk allocations and rewards, and often inexperienced owners and buyers find it hard to navigate an opaque marketplace. The result is poor project management and execution, insufficient skills, inadequate design processes, and underinvestment in skills development, R&D, and innovation.”
They describe seven areas where action is needed:
- reshape regulation and raise transparence, including streamlining permitting and approvals processes, reducing informality and corruption, and encouraging transparency on cost and performance; “Best practice regulation would include moving toward outcome-based, more standardized building codes, and consolidating land to promote scale.”
- rewire the contractual framework to reshape industry dynamics to encourage collaboration and problem solving and focus on best value and past performance rather than just cost;
- rethink design and engineering processes to incorporate value engineering with a focus on constructability and repeatable design elements;
- improve procurement and supply-chain management through digitally enabled processes;
- improve on-site execution through adoption of existing tools and processes to improve project planning and management, ensuring that all pre-work activity is complete before starting actual constriction, greater use of key performance indicators (KPIs), and adoption of lean operating principles;
- infuse digital technology, new materials, and advanced automation including increasing investment in IT and R&D, and appoint innovation officers to oversee technology adoption; and,
- reskill the workforce to be able to utilize the new technologies and processes.
The report goes on to note the possibility of major changes in construction:
“Parts of the industry could move toward a manufacturing-inspired mass-production system that would boost productivity up to tenfold. Industrial and infrastructure megaprojects need to instill holistic project-operating systems on-site and in design offices. The highly non-linear and challenging nature of megaprojects underscores the difficulty of, and necessity for, moving toward an industrialized project-operating system.”
The McKinsey report’s findings fit with my hypothesis of a deterioration of the 50s and 60s intangibles and a failure to renew the stock of intangible capital. The fact that their report highlights some of the same points made in my piece of decades ago is disconcerting. But it bolsters my argument. The construction industry done little to improve its production processes. It is little wonder that its productivity has declined.
At the end of his article, Klein admits that he has no ideas on how to increase construction productivity. I think the McKinsey report points the way to a solution. The trick is to get the companies and workers to implement the recommendation actions. There is one action I would highly recommend: require government agencies use their procurement policy to encourage innovation and new approaches by prescribing means and methods of delivery or requiring use of certain technologies. Mentioned only in passing in the follow up report by McKinsey, this suggestion deserves greater attention, including as part of the Biden Administration’s policy to use procurement to achieve environmental sustainability (Executive Order 14057).
In the end, however, it is up to the construction companies and workforce to take the actions need to increase productivity. To achieve this, the industry must accept responsibility for the problem. Rather than blame outside forces that are complicating the process, the industry (and policymakers) needs look inward. As Shakespeare told us, the fault is not in our stars but in ourselves.
The Intangible Economy 