(note: this post is more economics heavy than typical. I do not have any sort of training in economics, so it’s very possible I’ve made a mistake somewhere in interpreting this. If you see one, please point it out!) In general the evidence for long term lack of gains in construction efficiency is extensive enough that I haven’t felt compelled to dive into the specific measures documenting it. But (like with most things), the evidence isn’t so completely one-sided that any claims to the contrary can be immediately dismissed. And in general, the better we understand the various measures of construction efficiency, the better we can grasp the shape of the problem. So let’s take a look at a few of them.
I wonder if productivity is about to take an order of magnitude leap upwards. There are two major changes I see:
1. 3D printing is going large scale. Massive 3D printers can now print an entire house, though not all the fixtures, plumbing, wiring etc. But reports are coming in of houses printed in a few days that used to take weeks or months.
2. Factory-assembled modules, possibly using 3D printing indoors, may be about to really achieve economies of scale. The big impediment here has been the delivery constraints of the truck-bed, which is simply too narrow for what a crane could actually lift into place on the job site. It's the way highways and streets are made, for 2-passenger cars, that holds back this development aid. It would be interesting to see if areas where roads are not an issue - for example, on waterfront lots where barges can deliver huge modules - have the same issues. But even with this limit, it may become so much faster to construct multi-family modules in factories that this method adds great efficiency.
Also, it needs to be stressed that builders are being asked to comply with so many more codes, environmental goals (carbon neutral, etc.), and labor rules, that every gain in efficiency is probably clawed back by requirements in these areas. The famous Empire-State-Building-was-built-in-18-months example could never happen today.
I wonder whether there is any data / analysis on the regulatory impost on construction, which clearly has increased over time.. Also Health & Safety…. Compared to 50 years ago we have loads more H&S requirements and far more regulatory intervention… so that must have an effect…. Firstly on output cost and secondly on time…. So maybe the productivity delivered by innovation has been absorbed elsewhere so the net effect is close of zero….
Note that I am not decrying the H&S requirements etc…. It’s just that over time these have increased and its not clear whether this effect has been measured…. A useful target for research?
I wonder how labor costs would compare over time if you factored in tradesman turnover and illegal immigration. I'd think the trades have more turnover nowadays than before, increasing labor costs through hiring, training, and retaining knowledge. I'm not sure how you'd measure the use and numbers of illegal immigrant construction laborers, now vs before. My guess is illegal immigrant labor is subsidizing construction labor costs and the costs would actually be much higher with a more traditional workforce. The good thing is that with the rise of social media, many independent tradesmen are able to bid on jobs now, whereas before, these jobs were mostly given to bigger contracting companies (middlemen), who'd capture and retain more of this value.
I think a good common sense test is to look at how many people or how long it takes to construct a something today vs. 30 years ago. For example if productivity is growing at 2.4% a year, per the rule of 72 that means you should see the number of man-hours required to build a house to be cut in half over that period. Manufacturing industries with really meaningful productivity growth usually see big decreases in employment relative to output, for example GM employs half as many people or a quarter as many people today as it did a generation ago.
Based on what you've written in the past, it would be surprising if there have been recent meaningful gains in construction productivity. Over a long period of time, even 1% a year would be quite visible over a generation.
There might be a way to check it by looking at the annual report of a major publicly traded homebuilder today vs. 30 years ago. They have to report things like number of employees, costs, output., etc. One would have to have enough knowledge of the industry to know how to control for things like shifts between subcontracting and in-house, and geographical differences, but I think you can. I would be very surprised if analyzing it this way had a different result.
When I look at the data, I don't see low productivity growth as a big determinant of rising housing costs necessarily in recent years. If you look at the Case-Shiller price index for most cities, housing price growth has been fairly reasonable; true, it would be even lower with better productivity, but the actual monthly/annual cost of housing to the consumer is a function of both the upfront price and interest rates, and interest rates have gone way way down.
The bigger issue is you have the high cost cities where people want to move to, and there are a lot of zoning restrictions there. If a condo in New York costs $1,500 psf, even if you wave a magic wand and double productivity, maybe you're moving construction costs from $600 psf to $300, even that won't do a ton to move housing prices. The battle is mostly over the scarcity of development rights and taxes on development and so on. In fact I think even if you lower the cost of construction, you would just expect the value to get captured by developers/landowners - presumably at these prices, no one is holding back on construction because the economics don't pencil, so it won't stimulate more construction, so the benefits will just flow to the developers who capture the difference between the cost of construction and the sale price.
To second what Scott is saying below, the regulation impact seems non-trivial. Both in the sense that Scott points out and the temporal aspect of estimating these things. Off the top of my head I don't really know how you'd cope with it but off the cuff I'd think that if developers are seeing an increased wait time in permitting, EIS, etc, that would have significant costs associated and be impacting overall productivity. Both in the sense of having to pay for the labor to achieve the legal requirements and in the sense of delayed timelines it takes to produce their product.
On the economics front, you seemed to have done well. Perhaps a stronger treatment of the labor/capital trade off may have been beneficial but you quickly move into the calculus of economics when you really want to dig into that so I can understand why you'd avoid it for a general audience.
I wonder if productivity is about to take an order of magnitude leap upwards. There are two major changes I see:
1. 3D printing is going large scale. Massive 3D printers can now print an entire house, though not all the fixtures, plumbing, wiring etc. But reports are coming in of houses printed in a few days that used to take weeks or months.
2. Factory-assembled modules, possibly using 3D printing indoors, may be about to really achieve economies of scale. The big impediment here has been the delivery constraints of the truck-bed, which is simply too narrow for what a crane could actually lift into place on the job site. It's the way highways and streets are made, for 2-passenger cars, that holds back this development aid. It would be interesting to see if areas where roads are not an issue - for example, on waterfront lots where barges can deliver huge modules - have the same issues. But even with this limit, it may become so much faster to construct multi-family modules in factories that this method adds great efficiency.
Also, it needs to be stressed that builders are being asked to comply with so many more codes, environmental goals (carbon neutral, etc.), and labor rules, that every gain in efficiency is probably clawed back by requirements in these areas. The famous Empire-State-Building-was-built-in-18-months example could never happen today.
Hi, we'd love to have you write for Works in Progress. Please drop us a line hello (at) worksinprogress.co
I wonder whether there is any data / analysis on the regulatory impost on construction, which clearly has increased over time.. Also Health & Safety…. Compared to 50 years ago we have loads more H&S requirements and far more regulatory intervention… so that must have an effect…. Firstly on output cost and secondly on time…. So maybe the productivity delivered by innovation has been absorbed elsewhere so the net effect is close of zero….
Note that I am not decrying the H&S requirements etc…. It’s just that over time these have increased and its not clear whether this effect has been measured…. A useful target for research?
I wonder how labor costs would compare over time if you factored in tradesman turnover and illegal immigration. I'd think the trades have more turnover nowadays than before, increasing labor costs through hiring, training, and retaining knowledge. I'm not sure how you'd measure the use and numbers of illegal immigrant construction laborers, now vs before. My guess is illegal immigrant labor is subsidizing construction labor costs and the costs would actually be much higher with a more traditional workforce. The good thing is that with the rise of social media, many independent tradesmen are able to bid on jobs now, whereas before, these jobs were mostly given to bigger contracting companies (middlemen), who'd capture and retain more of this value.
I think a good common sense test is to look at how many people or how long it takes to construct a something today vs. 30 years ago. For example if productivity is growing at 2.4% a year, per the rule of 72 that means you should see the number of man-hours required to build a house to be cut in half over that period. Manufacturing industries with really meaningful productivity growth usually see big decreases in employment relative to output, for example GM employs half as many people or a quarter as many people today as it did a generation ago.
Based on what you've written in the past, it would be surprising if there have been recent meaningful gains in construction productivity. Over a long period of time, even 1% a year would be quite visible over a generation.
There might be a way to check it by looking at the annual report of a major publicly traded homebuilder today vs. 30 years ago. They have to report things like number of employees, costs, output., etc. One would have to have enough knowledge of the industry to know how to control for things like shifts between subcontracting and in-house, and geographical differences, but I think you can. I would be very surprised if analyzing it this way had a different result.
When I look at the data, I don't see low productivity growth as a big determinant of rising housing costs necessarily in recent years. If you look at the Case-Shiller price index for most cities, housing price growth has been fairly reasonable; true, it would be even lower with better productivity, but the actual monthly/annual cost of housing to the consumer is a function of both the upfront price and interest rates, and interest rates have gone way way down.
The bigger issue is you have the high cost cities where people want to move to, and there are a lot of zoning restrictions there. If a condo in New York costs $1,500 psf, even if you wave a magic wand and double productivity, maybe you're moving construction costs from $600 psf to $300, even that won't do a ton to move housing prices. The battle is mostly over the scarcity of development rights and taxes on development and so on. In fact I think even if you lower the cost of construction, you would just expect the value to get captured by developers/landowners - presumably at these prices, no one is holding back on construction because the economics don't pencil, so it won't stimulate more construction, so the benefits will just flow to the developers who capture the difference between the cost of construction and the sale price.
To second what Scott is saying below, the regulation impact seems non-trivial. Both in the sense that Scott points out and the temporal aspect of estimating these things. Off the top of my head I don't really know how you'd cope with it but off the cuff I'd think that if developers are seeing an increased wait time in permitting, EIS, etc, that would have significant costs associated and be impacting overall productivity. Both in the sense of having to pay for the labor to achieve the legal requirements and in the sense of delayed timelines it takes to produce their product.
On the economics front, you seemed to have done well. Perhaps a stronger treatment of the labor/capital trade off may have been beneficial but you quickly move into the calculus of economics when you really want to dig into that so I can understand why you'd avoid it for a general audience.