Last time in our series on the nature of the construction productivity problem, we looked at economies of scale, finding that capturing economies of scale in homebuilding is difficult.
Of all the levers we have to adjust housing cost, I’m convinced this is probably the least effective - because as you pointed out, we’ve already done a pretty good job with what we have. Short of some massive engineering or material breakthrough, any savings are likely to come from financing or permitting efficiency. Plus, if it costs 15% less to build, that doesn’t necessarily move the market, it just increases the developers margin.
It seems to be that with most innovations in construction the price is designed to match the alternative no matter the savings made by the manufacturer. Any savings are taken as extra margin and the product is sold for other reasons eg "these lighter elements will save you cost in the structure below ".
Is it that the construction market can never be as competitive since it's far less liquid than other markets? given the size of the stuff, the storage costs, the cap ex involved and given the relatively low cost to weight which makes transport significant.
Is it that construction is the biggest thing we make and therefore the hardest thing to benefit from standard economics?
I have no idea, but my guess is I think it's just because it's not really standard economics. You can't productize it because every site is different, it's heavily regulated, it's often but not always purchased with a 30-year-fixed-rate-loan(!) unique to this country, tons of subsidies on both sides of the transaction, long cycles. I would imagine this all trickles down to the components business. Like, what would happen if people just effectively stopped buying cars for 8 year periods and every county required a slightly different model?
I think w/r/t the risk averseness of building officials, ICC-ES's ESR program is supposed to bless these new construction products. However, I found that ICC-ES has about 80% market share in the construction space and outstanding high margins compared to the broader TIC (testing, inspection, and certification) industry. More here: https://moinmybackyard.substack.com/p/icc-es-the-financial-engine-of-the. In short, I think they are undersupplying the market for construction product certifications, thereby choking off innovation.
I hadn't looked at the margins of the construction product manufacturers like Simpson, and was surprised they are so large. Simpson is a major user of ICC-ES's ESRs. Part of their high margins, I think, come from the rarefied legitimacy that the ESR grants them.
Still, overall, I agree in the big picture, squeezing some margin out of one tiny little component of a building isn't going to move the needle on the overall problem. However, I do think ICC-ES and the broader TIC market does sit at a bottleneck above construction innovation in general. Unbottlenecking that may be a win.
There is one, albeit somewhat marginal, strategy for reducing material costs: volume. Unit costs of commodities tend to go down as the volume of purchases goes up. This is true of almost every commodity, construction materials included. Even if it's only a few percentage points, in a market as tightly strung as housing construction, that's nothing to sneeze at.
This is likely why more and more construction starts involve contractors that are placing orders for dozens, even hundreds of houses at once.
This is also a way of reducing labor costs, as it turns out. It's just cheaper to put up fifty houses in the same development at more-or-less the same time than it is to build those same fifty houses scattered across the county. Again, not a huge difference, but it's a big enough deal that contractors and subcontractors that do new construction generally don't do renovations/repairs/additions, and vice versa. The business model is too different.
As AI makes people more productive labor increases in price. So we might move towards automated systems but it won't reduce costs from existing methods, just to keep pace with where they are now. That might change material use patterns as automated systems will be better with some types of materials than others. Maybe even substituting more expensive materials if the total cost of installation is lower.
The thing I think might move the needle is much lower cost of transportation due to autonomy. Autonomous truck networks will want to have 100% utilization, and that will lower the price for loads that have fewer timing constraints, buffering the system keeping expensive capital in constant use. This might change some of the economics in this space, but it is very speculative.
I do wonder whether trucking automation/electrification will make moving manufactured components over longer distances more economic. Everything in this space is harder than you would expect, but it seems at least possible.
Given how squeezed truckers are today, I doubt autonomous trucks will make all that much difference. Truck drivers used to be fairly well paid. Aside from being on the road a fair bit, it was a job one could support a family on. Deregulation ended that and the labor costs, i.e. driver pay, collapsed. If anything, a move to autonomous trucks is likely to increase trucking costs since it will be controlled by a handful of large companies with little or no competition and trucking companies will be locked in to their suppliers by contracts and the DMCA.
Consider what Uber and Lyft did to the taxi system. Drivers earn less now and dispatching costs went from 10% in the old radio dispatch days to around 20% nowadays. Good luck getting around the dispatching oligopoly. Any efficiency obtained by eliminating truck drivers will turn into shareholder value not lower shipping costs.
Not true at all as consumer taxi costs are the equivalent here to building housing costs and those have definitely dropped down since Uber / Lyft arrived though at the expense of lower paid drivers
Consumer taxi costs were lower when Uber and Lyft were losing money. Now, Uber and Lyft are just as expensive as taxis whether on the meter or a flat rate, usually determined by checking what Uber or Lyft would charge.
I based my estimates on my incorrect memory of "I drove for Lyft for a week and learned its business model is broken". I had remembered 10% and 20%, but the numbers are more like 15% and 30%, that is before Uber and Lyft, taxi dispatch services cost drivers about 15% of what the passenger paid for the ride. Uber and Lyft now take closer to 30%. Consumers pay about the same, but drivers take home less than they used to. The difference is that Uber and Lyft dispatching services are more expensive than the old fashioned radio taxi model.
But they are worth it If you drove a radio taxi in the 80s or 90s, you'd know the amount of deadheading was like 3x of what it is now. Consumers pay less because longer rides are still likely to have a return fare, the market clears better with a larger network. There was also not fare adjustment for how busy it was. Additionally you had to work a shift, now you work when. you want. The cut the networks make is probably too low, but I was poor enough to remember my dad taking night shifts for taxi service. After his day job. So I remember what it was like. Imagine taking off because. you were sick and working on saturday instead. My dad couldn't d that and keep his extra job. but he could today.
That flexibility comes at a price. In Seattle, for example, half of all Uber/Lyft driving is deadhead. The drivers union argues that there are too many drivers and wants limits on the number active at any time. I know NYC limits the number of Uber/Lyft cars that can be active at any time to manage congestion. This is what happened with taxis in the 1930s. When anyone could enter the business as they chose, wages plummeted and congestion increased. Taxis might have been relatively cheap and cabs easy to find in high traffic areas, but wages were low, cabs poorly maintained and streets jammed with them.
The problem with Uber/Lyft is that they can't make a profit without taking a bigger cut of fares AND charging customers as much as a taxi would charge. When drivers bought a shift, they kept the fares after paying a fixed price. Now, Uber/Lyft takes as much as it can while still providing a ride, and their take is more than that of an old fashioned radio dispatch outfit. Now, drivers have flexible hours, but they get paid less and their pay is an unpredictable share of the customer fare.
I thought cheap labor was solved with immigrants? 40-50 years ago there were virtually no Mexican or Central American immigrants working in home construction. Now there are few projects that do not rely on immigrant labor. I'd think the manufactured home industry is similar. Construction costs are driven by policy choices and excessive profits in the FIRE sector.
I actually think there could be more to reducing the cost of materials than can be deduced by looking at the profit margins of material suppliers. Similarly to how Elon Musk initially got into SpaceX (what are the underlying component costs compared to the current delivered market price.)
Many house components can be reduced to a much more basic level than what is currently being used in the current "systems" paradigm. For instance a traditional window made from a small amount of oak wood, glass, and putty, from a material perspective is considerably cheaper than a factory window. Even a plaster interior wall is made from sand, lime, and hair. These components are much cheaper than buying drywall.
The trick, of course, is getting these materials assembled into a house. But historically it seems very possible to utilize skilled labor to perform this function. Look at how fast Victorian era cities were built using these base components and transforming them on site utilizing highly skilled guild labor.
This storyline has, by process of elimination, brought us closer to what the real answer is for reducing construction costs (housing construction costs). There is no way to appreciably lower housing construction costs until we choose to build in other forms and design paradigms. Folks keep thinking there's some secret way to reduce the price of building center entry colonials on subdivision cul de sacs. Nope.
I think we've hit the outer limits of what can be accomplished by building with dimensional lumber, and we need to think in new directions. This flies in the face of the fashion industry nature of American home building, i.e., every house being unique and expressing the inner dreams of its owner. Broad Corporation in Changsha has an interesting solution, although they've not been able to crack the American market due to all the usual regulatory and code restrictions.
I've been in construction for 50 years and started writing about it recently. I had a similar idea that construction could benefit from potential efficiencies and standarizations that wouldn't limit building flexibility. There had to be a way to reign in the approximatley 4X multiple on construction costs versus general inflation over the past 50 years. I did a fairly deep dive and came to the concluson that US construction (labor and materials) are already amazingly efficient. Builders know what they are doing and leverage labor well. Essentially all the excess housing cost inflation (I was mostly looking at housing, but other construction is similar, I think) results from non-direct cost increases: Fill-in housing versus large subdivisions, regulations, fees, buildable land scarcity, restrictive zoning and codes, just about everything other than the actual hard construction costs. True, wages have increased faster than inflation over the past 50 years, but so have efficiencies. What caused the cost of housing to go from 2-3X annual household income to 10-15X annual income is due to everything EXCEPT labor and materials. Some oversimplification, obviously, but improving building efficiency is not a viable solution to high housing costs.
The point about construction not being able to relocate to cheap labor matches what we see, but the industry runs the mirror image of it: instead of moving the work to the labor, it moves the labor to the work and pays a premium to do it. We run a job board for traveling trades workers, and on data center, refinery and plant jobs the going rate is roughly $38-45 an hour plus $100-125 a day per diem, which is a 20-30% add on top of the wage before overtime even starts. So the geographic wage differential does get arbitraged, just in the expensive direction - high-cost projects import crews instead of importing cheaper labor, and the premium is what buys the mobility. State-by-state numbers on wages and per diem if it is useful: https://roadtrades.com/pay-per-diem-index/
We have Artificial Intelligence, self driving vehicles and robots welding, soldering, packing, assembling parts, walking down streets with law enforcement, Robots building and assembling pre-fab homes, robot animals as mules and picking up our groceries! All while we play , build, create for ourselves.
Abundance has been slowly coming in. You just haven’t noticed it yet. Hang in there. Don’t worry you’re going to have the time of your life and money will not be an issue. 😉
FYI pay attention to what the billionaires and millionaires are saying, example Elon Musk, and the others. 💰💰💰
Worried about politics and red tape, why? Artificial intelligence will show you how we don’t need that and we don’t need large governments. If any the answers and solutions will be right in front of you in a nanosecond. The future is you won’t need to vote for somebody to make change, it will just get done. No more politics no more useless government control systems. Worried about work? you won’t need to, if you want to work It’s a choice because you enjoy it. Not because you have to.
Are the construction costs of buildings (per unit) lower than detached houses? It seems that same scale economy should apply here, although of course the technical complexity is higher.
The single biggest expense of building a home is the price of the land. If you want to get a building loan or just recruit limited partners to back a project, the cost of the land drives the overall financial structure, how much money is needed and the price of the unit to be sold. Some friends of mine build spec houses in my area, and it all starts with the lot.
Sometimes cost saving ideas fail to lower costs. Those metal plates that make it easier for framers were the enabling technology for McMansions. They made it possible to build larger houses with lower effective per square foot labor costs. It's like more efficient automobile engines. They could have led to an overall more efficient automotive fleet, but they wound up making SUVs less expensive to operate.
Also, the US, at least until recently, had the option of importing cheap labor. Roofers in California used to be unionized, and roofing paid well and offered middle class benefits. Immigrant labor was cheaper, and now most roofing in the state is done by immigrants who are paid much less and do without benefits. There's a big movement among conservatives to deregulate the licensed trades like plumbing and electrical. That could lower building costs, but good luck getting insurance.
Not the conclusion any of us wanted, but thank you for the step-by-step category walkthrough. We've seen very limited ability to cut costs. Many developers bring construction in-house, saving 4-5% on contractor fees. Meaningful, but it can be argued that even this savings is offset by the internalized risk assumed in the process.
Construction of multifamily housing units in the Northeast US is ~$275k/unit at the mid-range (secondary urban markets). This is significantly less than the cost to produce single family homes due to the cost of land, infrastructure and home size expectations. While not what some people want to hear, it is the answer - build smaller homes.
The familiarity issue can in some cases be solved by international consolidation, allowing trusted brands and distributors to leverage lower production costs. This works best for relatively high value, compact products, e.g fasteners, paint or electrical or plumbing fixtures.
There's also ongoing work on establishing design rules for less well known materials such as steel fibre reinforced concrete, which can reduce the need for labour-intensive rebar work.
Of all the levers we have to adjust housing cost, I’m convinced this is probably the least effective - because as you pointed out, we’ve already done a pretty good job with what we have. Short of some massive engineering or material breakthrough, any savings are likely to come from financing or permitting efficiency. Plus, if it costs 15% less to build, that doesn’t necessarily move the market, it just increases the developers margin.
It seems to be that with most innovations in construction the price is designed to match the alternative no matter the savings made by the manufacturer. Any savings are taken as extra margin and the product is sold for other reasons eg "these lighter elements will save you cost in the structure below ".
Is it that the construction market can never be as competitive since it's far less liquid than other markets? given the size of the stuff, the storage costs, the cap ex involved and given the relatively low cost to weight which makes transport significant.
Is it that construction is the biggest thing we make and therefore the hardest thing to benefit from standard economics?
I have no idea, but my guess is I think it's just because it's not really standard economics. You can't productize it because every site is different, it's heavily regulated, it's often but not always purchased with a 30-year-fixed-rate-loan(!) unique to this country, tons of subsidies on both sides of the transaction, long cycles. I would imagine this all trickles down to the components business. Like, what would happen if people just effectively stopped buying cars for 8 year periods and every county required a slightly different model?
I think w/r/t the risk averseness of building officials, ICC-ES's ESR program is supposed to bless these new construction products. However, I found that ICC-ES has about 80% market share in the construction space and outstanding high margins compared to the broader TIC (testing, inspection, and certification) industry. More here: https://moinmybackyard.substack.com/p/icc-es-the-financial-engine-of-the. In short, I think they are undersupplying the market for construction product certifications, thereby choking off innovation.
I hadn't looked at the margins of the construction product manufacturers like Simpson, and was surprised they are so large. Simpson is a major user of ICC-ES's ESRs. Part of their high margins, I think, come from the rarefied legitimacy that the ESR grants them.
Still, overall, I agree in the big picture, squeezing some margin out of one tiny little component of a building isn't going to move the needle on the overall problem. However, I do think ICC-ES and the broader TIC market does sit at a bottleneck above construction innovation in general. Unbottlenecking that may be a win.
There is one, albeit somewhat marginal, strategy for reducing material costs: volume. Unit costs of commodities tend to go down as the volume of purchases goes up. This is true of almost every commodity, construction materials included. Even if it's only a few percentage points, in a market as tightly strung as housing construction, that's nothing to sneeze at.
This is likely why more and more construction starts involve contractors that are placing orders for dozens, even hundreds of houses at once.
This is also a way of reducing labor costs, as it turns out. It's just cheaper to put up fifty houses in the same development at more-or-less the same time than it is to build those same fifty houses scattered across the county. Again, not a huge difference, but it's a big enough deal that contractors and subcontractors that do new construction generally don't do renovations/repairs/additions, and vice versa. The business model is too different.
As AI makes people more productive labor increases in price. So we might move towards automated systems but it won't reduce costs from existing methods, just to keep pace with where they are now. That might change material use patterns as automated systems will be better with some types of materials than others. Maybe even substituting more expensive materials if the total cost of installation is lower.
The thing I think might move the needle is much lower cost of transportation due to autonomy. Autonomous truck networks will want to have 100% utilization, and that will lower the price for loads that have fewer timing constraints, buffering the system keeping expensive capital in constant use. This might change some of the economics in this space, but it is very speculative.
I do wonder whether trucking automation/electrification will make moving manufactured components over longer distances more economic. Everything in this space is harder than you would expect, but it seems at least possible.
Given how squeezed truckers are today, I doubt autonomous trucks will make all that much difference. Truck drivers used to be fairly well paid. Aside from being on the road a fair bit, it was a job one could support a family on. Deregulation ended that and the labor costs, i.e. driver pay, collapsed. If anything, a move to autonomous trucks is likely to increase trucking costs since it will be controlled by a handful of large companies with little or no competition and trucking companies will be locked in to their suppliers by contracts and the DMCA.
Consider what Uber and Lyft did to the taxi system. Drivers earn less now and dispatching costs went from 10% in the old radio dispatch days to around 20% nowadays. Good luck getting around the dispatching oligopoly. Any efficiency obtained by eliminating truck drivers will turn into shareholder value not lower shipping costs.
Not true at all as consumer taxi costs are the equivalent here to building housing costs and those have definitely dropped down since Uber / Lyft arrived though at the expense of lower paid drivers
Consumer taxi costs were lower when Uber and Lyft were losing money. Now, Uber and Lyft are just as expensive as taxis whether on the meter or a flat rate, usually determined by checking what Uber or Lyft would charge.
I based my estimates on my incorrect memory of "I drove for Lyft for a week and learned its business model is broken". I had remembered 10% and 20%, but the numbers are more like 15% and 30%, that is before Uber and Lyft, taxi dispatch services cost drivers about 15% of what the passenger paid for the ride. Uber and Lyft now take closer to 30%. Consumers pay about the same, but drivers take home less than they used to. The difference is that Uber and Lyft dispatching services are more expensive than the old fashioned radio taxi model.
https://www.fullstackeconomics.com/p/i-drove-for-lyft-for-a-week-and-learned
But they are worth it If you drove a radio taxi in the 80s or 90s, you'd know the amount of deadheading was like 3x of what it is now. Consumers pay less because longer rides are still likely to have a return fare, the market clears better with a larger network. There was also not fare adjustment for how busy it was. Additionally you had to work a shift, now you work when. you want. The cut the networks make is probably too low, but I was poor enough to remember my dad taking night shifts for taxi service. After his day job. So I remember what it was like. Imagine taking off because. you were sick and working on saturday instead. My dad couldn't d that and keep his extra job. but he could today.
That flexibility comes at a price. In Seattle, for example, half of all Uber/Lyft driving is deadhead. The drivers union argues that there are too many drivers and wants limits on the number active at any time. I know NYC limits the number of Uber/Lyft cars that can be active at any time to manage congestion. This is what happened with taxis in the 1930s. When anyone could enter the business as they chose, wages plummeted and congestion increased. Taxis might have been relatively cheap and cabs easy to find in high traffic areas, but wages were low, cabs poorly maintained and streets jammed with them.
The problem with Uber/Lyft is that they can't make a profit without taking a bigger cut of fares AND charging customers as much as a taxi would charge. When drivers bought a shift, they kept the fares after paying a fixed price. Now, Uber/Lyft takes as much as it can while still providing a ride, and their take is more than that of an old fashioned radio dispatch outfit. Now, drivers have flexible hours, but they get paid less and their pay is an unpredictable share of the customer fare.
I thought cheap labor was solved with immigrants? 40-50 years ago there were virtually no Mexican or Central American immigrants working in home construction. Now there are few projects that do not rely on immigrant labor. I'd think the manufactured home industry is similar. Construction costs are driven by policy choices and excessive profits in the FIRE sector.
Construction labor has always been solved with immigrants.
I actually think there could be more to reducing the cost of materials than can be deduced by looking at the profit margins of material suppliers. Similarly to how Elon Musk initially got into SpaceX (what are the underlying component costs compared to the current delivered market price.)
Many house components can be reduced to a much more basic level than what is currently being used in the current "systems" paradigm. For instance a traditional window made from a small amount of oak wood, glass, and putty, from a material perspective is considerably cheaper than a factory window. Even a plaster interior wall is made from sand, lime, and hair. These components are much cheaper than buying drywall.
The trick, of course, is getting these materials assembled into a house. But historically it seems very possible to utilize skilled labor to perform this function. Look at how fast Victorian era cities were built using these base components and transforming them on site utilizing highly skilled guild labor.
This storyline has, by process of elimination, brought us closer to what the real answer is for reducing construction costs (housing construction costs). There is no way to appreciably lower housing construction costs until we choose to build in other forms and design paradigms. Folks keep thinking there's some secret way to reduce the price of building center entry colonials on subdivision cul de sacs. Nope.
I think we've hit the outer limits of what can be accomplished by building with dimensional lumber, and we need to think in new directions. This flies in the face of the fashion industry nature of American home building, i.e., every house being unique and expressing the inner dreams of its owner. Broad Corporation in Changsha has an interesting solution, although they've not been able to crack the American market due to all the usual regulatory and code restrictions.
I've been in construction for 50 years and started writing about it recently. I had a similar idea that construction could benefit from potential efficiencies and standarizations that wouldn't limit building flexibility. There had to be a way to reign in the approximatley 4X multiple on construction costs versus general inflation over the past 50 years. I did a fairly deep dive and came to the concluson that US construction (labor and materials) are already amazingly efficient. Builders know what they are doing and leverage labor well. Essentially all the excess housing cost inflation (I was mostly looking at housing, but other construction is similar, I think) results from non-direct cost increases: Fill-in housing versus large subdivisions, regulations, fees, buildable land scarcity, restrictive zoning and codes, just about everything other than the actual hard construction costs. True, wages have increased faster than inflation over the past 50 years, but so have efficiencies. What caused the cost of housing to go from 2-3X annual household income to 10-15X annual income is due to everything EXCEPT labor and materials. Some oversimplification, obviously, but improving building efficiency is not a viable solution to high housing costs.
The point about construction not being able to relocate to cheap labor matches what we see, but the industry runs the mirror image of it: instead of moving the work to the labor, it moves the labor to the work and pays a premium to do it. We run a job board for traveling trades workers, and on data center, refinery and plant jobs the going rate is roughly $38-45 an hour plus $100-125 a day per diem, which is a 20-30% add on top of the wage before overtime even starts. So the geographic wage differential does get arbitraged, just in the expensive direction - high-cost projects import crews instead of importing cheaper labor, and the premium is what buys the mobility. State-by-state numbers on wages and per diem if it is useful: https://roadtrades.com/pay-per-diem-index/
We have Artificial Intelligence, self driving vehicles and robots welding, soldering, packing, assembling parts, walking down streets with law enforcement, Robots building and assembling pre-fab homes, robot animals as mules and picking up our groceries! All while we play , build, create for ourselves.
Abundance has been slowly coming in. You just haven’t noticed it yet. Hang in there. Don’t worry you’re going to have the time of your life and money will not be an issue. 😉
FYI pay attention to what the billionaires and millionaires are saying, example Elon Musk, and the others. 💰💰💰
Worried about politics and red tape, why? Artificial intelligence will show you how we don’t need that and we don’t need large governments. If any the answers and solutions will be right in front of you in a nanosecond. The future is you won’t need to vote for somebody to make change, it will just get done. No more politics no more useless government control systems. Worried about work? you won’t need to, if you want to work It’s a choice because you enjoy it. Not because you have to.
Enjoy ❤️🤖🦾
Are the construction costs of buildings (per unit) lower than detached houses? It seems that same scale economy should apply here, although of course the technical complexity is higher.
The single biggest expense of building a home is the price of the land. If you want to get a building loan or just recruit limited partners to back a project, the cost of the land drives the overall financial structure, how much money is needed and the price of the unit to be sold. Some friends of mine build spec houses in my area, and it all starts with the lot.
Sometimes cost saving ideas fail to lower costs. Those metal plates that make it easier for framers were the enabling technology for McMansions. They made it possible to build larger houses with lower effective per square foot labor costs. It's like more efficient automobile engines. They could have led to an overall more efficient automotive fleet, but they wound up making SUVs less expensive to operate.
Also, the US, at least until recently, had the option of importing cheap labor. Roofers in California used to be unionized, and roofing paid well and offered middle class benefits. Immigrant labor was cheaper, and now most roofing in the state is done by immigrants who are paid much less and do without benefits. There's a big movement among conservatives to deregulate the licensed trades like plumbing and electrical. That could lower building costs, but good luck getting insurance.
Not the conclusion any of us wanted, but thank you for the step-by-step category walkthrough. We've seen very limited ability to cut costs. Many developers bring construction in-house, saving 4-5% on contractor fees. Meaningful, but it can be argued that even this savings is offset by the internalized risk assumed in the process.
Construction of multifamily housing units in the Northeast US is ~$275k/unit at the mid-range (secondary urban markets). This is significantly less than the cost to produce single family homes due to the cost of land, infrastructure and home size expectations. While not what some people want to hear, it is the answer - build smaller homes.
We can allow more guest workers and eliminate tariffs as a first step.
The familiarity issue can in some cases be solved by international consolidation, allowing trusted brands and distributors to leverage lower production costs. This works best for relatively high value, compact products, e.g fasteners, paint or electrical or plumbing fixtures.
There's also ongoing work on establishing design rules for less well known materials such as steel fibre reinforced concrete, which can reduce the need for labour-intensive rebar work.