In the New York Times, Ezra Klein investigated the recent Goolsbee and Syverson paper on construction productivity we recently looked at. Klein suggests that the stagnation in construction productivity might be the result of organized special interests increasingly leveraging their influence over various parts of the construction process. This shows up as more complex and difficult to accommodate regulations, and groups whose approval must be secured (often at great expense) to get something built. While all sectors display this to some extent, construction, by virtue of the number of stakeholders involved (local residents, politicians, zoning boards, etc.) might show it more than others. Here’s Klein:
“My point is that to be an explanation for the productivity stagnation in construction, regulation needs to be a much greater burden on construction than it is in other sectors, and it’s not obvious that that’s the case.”
This assumes the cost of complying with regulation is same for construction as for other industries which I do not think is the case. For a manufactured good or an large online service, the cost of figuring out how to comply is amortized over 1M or even 100M units. The impact to the bill of material is still paid by every unit but the substantial costs of understanding the regulation and engineering a solution are not as they are amortized over the whole production. For construction, the cost of understanding and figuring out how to comply has to be done for every single building.
Regulation compliance is absolutely harder for MF construction than for single family homes. Furthermore, single family homes have an inbuilt advantage of being more generally acceptable, at least in neighborhoods already zoned for them, which is more often the case than for MF construction, which often requires multiple layers of approval (here in New York City, that means the dreaded ULURP - Uniform Land Use Review Process - which can take a year or more, and pre-ULURP, which is basically unlimited, or until somebody's will breaks). In New York City, a major set of LSRD (Large Scale Residential Development) buildings was already at the foundation laying stage when it was challenged by a new New York Constitutional amendment guaranteeing clean air and water. This amendment wasn't voted in by referendum until AFTER the foundations were being laid, so neighborhood groups convince the judge to suspend construction while they argued the projects (3 towers, 4 developers) would violate the new amendment. While this delay takes place, the developers are losing millions, and one has already sold the lot to another developer (i.e. their will was broken).
Extra regulation is inevitable to some degree in a growing population, especially in cities - which residents of SF home communities fight like dogs not to join, in spite of the inefficiencies of suburbs over time - but when almost all groups are NIMBY, including many of the politicians elected, it's hard to see how anything developers or construction professionals can do, can counter that.
Also, even SF homes contain the result of countless regulations going into every component and facet of a home. Not all of this is bad: today's homes come with built-ins that weren't available in the 1960s, as well as safer electricals, plumbing, etc. (also, MORE outlets for all the stuff we plug in now). But it seems like some things have actually gotten worse: hollow doors instead of solid wooden doors, oversized windows and column-free rooms, weakening homes in the event of storms/hurricanes (are these more frequent due to climate change now, or are houses just built worse now? This documentary makes a case for the latter: https://www.thelasthousestanding.org/).
No one would take the chances workers took when they built the Empire State Building in 18 months, today, and such towers are both more complex and more regulated. It's certainly more than 0.5% extra time, when you factor in all the supply chained components.
I see there was a very sharp dropoff in regulations in your Figure 5 about the Federal Registry, in 2016, at the very end of the measured time period. Did construction productivity jump after that?
I am beginning to wonder if there’s really a problem to solve here. Baumol’s Cost Law is pretty damned ironclad in a developed economy; it drives basically all wage increases outside highly automated industries like manufacturing or agriculture and new high-value-added ones like software development.
The innate value of an hour-long massage is unknowable; it exists only in the context of how much disposable income sloshes around a society after it meets other needs.
For construction labor, prevailing wages are set by the need to compete with manufacturing work requiring similar work ethic and skills, but which has vastly increased in per-wage unit productivity for a century and a half.
Short of automating construction work, which I now very much believe is a pipe dream, having seen successive silver bullets failed to live up to their mooted potential, there’s no way out of this.
Those regulations accrete for a variety of reasons.
Some years back, I audited the code for the flight planning system for a major airline. They wanted to rewrite the code which was developed in an extinct version of 1960s FORTRAN and ran on a long obsolete computer. The code was full of kludges and special cases, some involving obscure database flags that modified various parameters in various ways. Some of the code checked for specific airports, aircraft types, deferred maintenance items and so on. I asked our informant at the airline, and he would explain that this code was in response to a crash in Hong Kong. Those few lines were a wind shear incident that knocked a plane off the runway during landing. That section about engine thrust was about an agreement with the pilot's union and the FAA. Regulations are like this. When something fails, the parties involved, the government and insurers sit down and figure out what went wrong and come up with something to fix it.
Similarly, the increased size of regulation can be about trying to eliminate ambiguities in the code that could be exploited by the unscrupulous. I remember some congressman complaining that the army had 20 pages of regulations describing a cherry pie. Sounds hilarious. The story was an indictment of our private sector. Contractors were supplying cherry pies with no cherries or with no lower crust. Contractors were cutting the size and mass of the pies, raising the crust to filling ratio and otherwise cutting costs while still satisfying the letter of the law but not satisfying anyone who might want to eat cherry pie. So, the regulations grew to specify crust size, crust thickness, cherry count, the definition of a cherry for the purposes of providing a cherry pie and so on.
Those building codes aren't growing because people just love adding restrictions and raising costs. They are growing because houses have burned down, wiring has burned out, people got cancer, energy costs were soaring, roofs went soaring and the so on. Sometimes, they grew because of new building materials. A lot of newer buildings used compressed wood products. The new ones are much more advanced than simple plywood and are used in structural components. I am sure that there are dozens of pages prescribing their composition, their resistance to fire, their strength and so on. Those regulations are a good thing in that they codify how to use a new material and actually make it easier to do so.
They used to say that no one whines like a farmer, but now as we are a less agricultural nation, no one whines like a businessman and the wealthier, the whinier.
I think the big cost driver in construction is the rising price of land. When land was cheap, one could build a cheap house on it. Now that land is expensive, one has to put a more expensive house on it. This is driven by both lenders and purchasers. Even if you wanted a $100K house on your $200K plot of land, you couldn't get financing for it. Houses have bigger budgets, so builders add amenities, upgrade materials, add space and provide more conveniences to meet them.
Projects can be impacted in different ways. As a local government capital improvements project manager, my CIP projects can sometimes face a heavy burden of collaboration with 3rd parties which can delay projects by a significant period of time. The cost impacts an often be consulting costs.
Federally funded projects have many burdens like historical and archeological studies which are often on-sensical and also slow down project delivery. I would prefer projects just pay historical agencies a direct fee and flip flop the burden of legal compliance. There are projects that just don’t get built because of something of marginal historical value.
On the regulatory front - it’d be worth exploring one extra thing that I’ve generally only encountered in the construction/infrastructure world: incorporation by reference. Whether it is the Federal Register/CFR, the building code, or zoning docs - they almost always incorporate myriad other requirements by reference: ASCE, ACI, ASHRAE, NSF, etc. not saying that means it is “the cause” but makes plausible that there is more of a regulatory impact than some other sectors.
One interesting thing I’m also thinking of (and applaud you for looking at single family) is the pervasiveness of federal dollars and therefore “buy America”, wage requirements, etc that may not otherwise apply. Of course, those are likely to NOT be an issue on the single family side.
Labor productivity per 100 sq ft is largely unchanged, and materials used for home construction are also largely unchanged. So what has changed? The size of home, from 1,695 sf in 1973 to 2,881 in 2017 (source: https://www.census.gov/construction/chars/pdf/c25ann2017.pdf , page 347). Increased floor area appears to be the source of higher single family home costs.
Also, in your IRC chart, by converting to 2021 dollars, you are counting inflation twice: once for code mandated construction cost increases, and once for overall inflation.
Seen this way, single family home construction costs are largely unchanged. As you note, soft costs and veto points like planning approvals, CEQA reviews (in California), developer fees etc. have increased.
For other forms of construction, anecdotally there is a gap between public sector construction costs and private sector construction costs. Pure speculation: the public sector over time has reduced its capacity to administer construction projects (relying on consultants instead), and has developed a reputation for avoiding litigation.
Hired consultants often have incentives that do not align with the public’s (e.g. a construction management firm may be a subsidiary of a larger contracting firm). Contractor claims for additional construction costs, instead of being contested, are justified by construction managers as ‘unknown conditions’, ‘design errors’, etc. And contractors often take an aggressive approach, knowing that public entities are reluctant to litigate. (It is common contractor practice in public sector construction to underbid a project so you are awarded the contract, and then make your profit via change orders.)
But in both public and private construction, soft costs and veto points have increased. With rising interest rates, these costs will assume still greater significance.
Insightful analysis as always, Brian. Thank you. I work in healthcare construction and 'feels' like it parallels your article of cost increases in the nuclear industries very well. A few thoughts come to mind:
1) Are some sectors disproportionally affecting overall productivity numbers? E.g. is healthcare dragging down the numbers that won't necessarily show up in single family?
2) I am just realizing that we are only talking about on-site productivity. It is terrifying to think about how much worse productivity would be if the whole cost of the project (mainly including professional services drug out by extended timelines) were included.
Also, other industries have a greater array of tools at their disposal to either evade or shift the costs of regulatory burden around. Moving manufacturing overseas, changing materials, redesigning products wholesale from scratch, or shifting production focus into more profitable categories (ie, shifting from sedans to minivans.) There's just less flexibility overall in construction to outsource or change housing designs or materials used since moving off of an industry standard increases the odds of running afoul of code or in some cases may even be illegal entirely.
The simple fact that SFH construction is so prevalent is an argument that you’re underestimating the impact here: everywhere else, SFH’s are comparatively rare because the cost of building (and then living in) an apartment is so much lower.
"Construction is an essential sector of the economy, responsible for building and maintaining the physical infrastructure that underpins our society. However, it’s no secret that construction productivity lags behind other sectors of the economy, such as manufacturing and information technology. So why is this the case?
One of the main reasons for the productivity gap is the unique nature of the construction industry. Unlike other sectors, construction projects are often one-off, bespoke endeavors, making it challenging to achieve the economies of scale that are typical of manufacturing or technology. Each project requires a different set of skills, tools, and materials, which can be costly and time-consuming to source and manage. This leads to a lack of standardization and efficiency, which can hinder productivity."
I have only a superficial knowledge, but I'm not aware of any innovations in home building in the last 50 years that ought to have produced big productivity gains. Manufactured homes have never really taken off, mainly because they haven't overcome the problem of rapid depreciation, AFAICT.
“My point is that to be an explanation for the productivity stagnation in construction, regulation needs to be a much greater burden on construction than it is in other sectors, and it’s not obvious that that’s the case.”
This assumes the cost of complying with regulation is same for construction as for other industries which I do not think is the case. For a manufactured good or an large online service, the cost of figuring out how to comply is amortized over 1M or even 100M units. The impact to the bill of material is still paid by every unit but the substantial costs of understanding the regulation and engineering a solution are not as they are amortized over the whole production. For construction, the cost of understanding and figuring out how to comply has to be done for every single building.
Regulation compliance is absolutely harder for MF construction than for single family homes. Furthermore, single family homes have an inbuilt advantage of being more generally acceptable, at least in neighborhoods already zoned for them, which is more often the case than for MF construction, which often requires multiple layers of approval (here in New York City, that means the dreaded ULURP - Uniform Land Use Review Process - which can take a year or more, and pre-ULURP, which is basically unlimited, or until somebody's will breaks). In New York City, a major set of LSRD (Large Scale Residential Development) buildings was already at the foundation laying stage when it was challenged by a new New York Constitutional amendment guaranteeing clean air and water. This amendment wasn't voted in by referendum until AFTER the foundations were being laid, so neighborhood groups convince the judge to suspend construction while they argued the projects (3 towers, 4 developers) would violate the new amendment. While this delay takes place, the developers are losing millions, and one has already sold the lot to another developer (i.e. their will was broken).
Extra regulation is inevitable to some degree in a growing population, especially in cities - which residents of SF home communities fight like dogs not to join, in spite of the inefficiencies of suburbs over time - but when almost all groups are NIMBY, including many of the politicians elected, it's hard to see how anything developers or construction professionals can do, can counter that.
Also, even SF homes contain the result of countless regulations going into every component and facet of a home. Not all of this is bad: today's homes come with built-ins that weren't available in the 1960s, as well as safer electricals, plumbing, etc. (also, MORE outlets for all the stuff we plug in now). But it seems like some things have actually gotten worse: hollow doors instead of solid wooden doors, oversized windows and column-free rooms, weakening homes in the event of storms/hurricanes (are these more frequent due to climate change now, or are houses just built worse now? This documentary makes a case for the latter: https://www.thelasthousestanding.org/).
No one would take the chances workers took when they built the Empire State Building in 18 months, today, and such towers are both more complex and more regulated. It's certainly more than 0.5% extra time, when you factor in all the supply chained components.
I see there was a very sharp dropoff in regulations in your Figure 5 about the Federal Registry, in 2016, at the very end of the measured time period. Did construction productivity jump after that?
I am beginning to wonder if there’s really a problem to solve here. Baumol’s Cost Law is pretty damned ironclad in a developed economy; it drives basically all wage increases outside highly automated industries like manufacturing or agriculture and new high-value-added ones like software development.
The innate value of an hour-long massage is unknowable; it exists only in the context of how much disposable income sloshes around a society after it meets other needs.
For construction labor, prevailing wages are set by the need to compete with manufacturing work requiring similar work ethic and skills, but which has vastly increased in per-wage unit productivity for a century and a half.
Short of automating construction work, which I now very much believe is a pipe dream, having seen successive silver bullets failed to live up to their mooted potential, there’s no way out of this.
Those regulations accrete for a variety of reasons.
Some years back, I audited the code for the flight planning system for a major airline. They wanted to rewrite the code which was developed in an extinct version of 1960s FORTRAN and ran on a long obsolete computer. The code was full of kludges and special cases, some involving obscure database flags that modified various parameters in various ways. Some of the code checked for specific airports, aircraft types, deferred maintenance items and so on. I asked our informant at the airline, and he would explain that this code was in response to a crash in Hong Kong. Those few lines were a wind shear incident that knocked a plane off the runway during landing. That section about engine thrust was about an agreement with the pilot's union and the FAA. Regulations are like this. When something fails, the parties involved, the government and insurers sit down and figure out what went wrong and come up with something to fix it.
Similarly, the increased size of regulation can be about trying to eliminate ambiguities in the code that could be exploited by the unscrupulous. I remember some congressman complaining that the army had 20 pages of regulations describing a cherry pie. Sounds hilarious. The story was an indictment of our private sector. Contractors were supplying cherry pies with no cherries or with no lower crust. Contractors were cutting the size and mass of the pies, raising the crust to filling ratio and otherwise cutting costs while still satisfying the letter of the law but not satisfying anyone who might want to eat cherry pie. So, the regulations grew to specify crust size, crust thickness, cherry count, the definition of a cherry for the purposes of providing a cherry pie and so on.
Those building codes aren't growing because people just love adding restrictions and raising costs. They are growing because houses have burned down, wiring has burned out, people got cancer, energy costs were soaring, roofs went soaring and the so on. Sometimes, they grew because of new building materials. A lot of newer buildings used compressed wood products. The new ones are much more advanced than simple plywood and are used in structural components. I am sure that there are dozens of pages prescribing their composition, their resistance to fire, their strength and so on. Those regulations are a good thing in that they codify how to use a new material and actually make it easier to do so.
They used to say that no one whines like a farmer, but now as we are a less agricultural nation, no one whines like a businessman and the wealthier, the whinier.
I think the big cost driver in construction is the rising price of land. When land was cheap, one could build a cheap house on it. Now that land is expensive, one has to put a more expensive house on it. This is driven by both lenders and purchasers. Even if you wanted a $100K house on your $200K plot of land, you couldn't get financing for it. Houses have bigger budgets, so builders add amenities, upgrade materials, add space and provide more conveniences to meet them.
Projects can be impacted in different ways. As a local government capital improvements project manager, my CIP projects can sometimes face a heavy burden of collaboration with 3rd parties which can delay projects by a significant period of time. The cost impacts an often be consulting costs.
Federally funded projects have many burdens like historical and archeological studies which are often on-sensical and also slow down project delivery. I would prefer projects just pay historical agencies a direct fee and flip flop the burden of legal compliance. There are projects that just don’t get built because of something of marginal historical value.
Truly excellent coverage here. The only question I have left is, how does this look in other countries? (e.g. is Japan facing the same trends?)
On the regulatory front - it’d be worth exploring one extra thing that I’ve generally only encountered in the construction/infrastructure world: incorporation by reference. Whether it is the Federal Register/CFR, the building code, or zoning docs - they almost always incorporate myriad other requirements by reference: ASCE, ACI, ASHRAE, NSF, etc. not saying that means it is “the cause” but makes plausible that there is more of a regulatory impact than some other sectors.
One interesting thing I’m also thinking of (and applaud you for looking at single family) is the pervasiveness of federal dollars and therefore “buy America”, wage requirements, etc that may not otherwise apply. Of course, those are likely to NOT be an issue on the single family side.
Labor productivity per 100 sq ft is largely unchanged, and materials used for home construction are also largely unchanged. So what has changed? The size of home, from 1,695 sf in 1973 to 2,881 in 2017 (source: https://www.census.gov/construction/chars/pdf/c25ann2017.pdf , page 347). Increased floor area appears to be the source of higher single family home costs.
Also, in your IRC chart, by converting to 2021 dollars, you are counting inflation twice: once for code mandated construction cost increases, and once for overall inflation.
Seen this way, single family home construction costs are largely unchanged. As you note, soft costs and veto points like planning approvals, CEQA reviews (in California), developer fees etc. have increased.
For other forms of construction, anecdotally there is a gap between public sector construction costs and private sector construction costs. Pure speculation: the public sector over time has reduced its capacity to administer construction projects (relying on consultants instead), and has developed a reputation for avoiding litigation.
Hired consultants often have incentives that do not align with the public’s (e.g. a construction management firm may be a subsidiary of a larger contracting firm). Contractor claims for additional construction costs, instead of being contested, are justified by construction managers as ‘unknown conditions’, ‘design errors’, etc. And contractors often take an aggressive approach, knowing that public entities are reluctant to litigate. (It is common contractor practice in public sector construction to underbid a project so you are awarded the contract, and then make your profit via change orders.)
But in both public and private construction, soft costs and veto points have increased. With rising interest rates, these costs will assume still greater significance.
Building buildings is artesanal work. Isn't that the gist of the problem?
Call TSMC's Arizona plant project manager for some insights.
His boss, Morris Chang, claims that its construction cost 4x more and takes 2x longer than in Taiwan – with no gain in quality.
Looks like we have a mystery!
What do you think the most likely causes of the slowdown n productivity are?
Insightful analysis as always, Brian. Thank you. I work in healthcare construction and 'feels' like it parallels your article of cost increases in the nuclear industries very well. A few thoughts come to mind:
1) Are some sectors disproportionally affecting overall productivity numbers? E.g. is healthcare dragging down the numbers that won't necessarily show up in single family?
2) I am just realizing that we are only talking about on-site productivity. It is terrifying to think about how much worse productivity would be if the whole cost of the project (mainly including professional services drug out by extended timelines) were included.
Also, other industries have a greater array of tools at their disposal to either evade or shift the costs of regulatory burden around. Moving manufacturing overseas, changing materials, redesigning products wholesale from scratch, or shifting production focus into more profitable categories (ie, shifting from sedans to minivans.) There's just less flexibility overall in construction to outsource or change housing designs or materials used since moving off of an industry standard increases the odds of running afoul of code or in some cases may even be illegal entirely.
The simple fact that SFH construction is so prevalent is an argument that you’re underestimating the impact here: everywhere else, SFH’s are comparatively rare because the cost of building (and then living in) an apartment is so much lower.
Developer Brandon Donnelly posted in his latest blog an answer to why productivity lags in the construction industry too: https://brandondonnelly.com/2023/02/18/why-construction-productivity-lags-other-sectors-of-the-economy/
"Construction is an essential sector of the economy, responsible for building and maintaining the physical infrastructure that underpins our society. However, it’s no secret that construction productivity lags behind other sectors of the economy, such as manufacturing and information technology. So why is this the case?
One of the main reasons for the productivity gap is the unique nature of the construction industry. Unlike other sectors, construction projects are often one-off, bespoke endeavors, making it challenging to achieve the economies of scale that are typical of manufacturing or technology. Each project requires a different set of skills, tools, and materials, which can be costly and time-consuming to source and manage. This leads to a lack of standardization and efficiency, which can hinder productivity."
Follow the link for the rest.
Can you tell it was written by ChatGPT?
I have only a superficial knowledge, but I'm not aware of any innovations in home building in the last 50 years that ought to have produced big productivity gains. Manufactured homes have never really taken off, mainly because they haven't overcome the problem of rapid depreciation, AFAICT.