Upon studying the housing crisis in Ireland for a while, there’s a depressing realisation that everything comes back to construction costs. 

Why are we reliant on global pension funds to build apartment blocks? Construction costs are too high for anyone else to do it. Why, this time around, have we not seen huge housing development on the scale of the Celtic Tiger? Construction costs. Why has the government struggled to build social housing? Construction costs. Why, simply, is Ireland unaffordable? Construction costs.

There’s an important chart in the Central Bank of Ireland’s 2019 report into the Irish housing market. I’ve reproduced it below. It shows the relationship between the number of houses built in a given year, and house prices. The pink dots are from 2010 to 2018, the blue ones re from 1995-2007 (we can ignore the green ones for now).

As you can see, the relationship between house prices and supply changed between 2007 and 2010. After 2010, for a given level of house prices, much fewer houses were built than in the 1995-2007 period.

Look at the gap between the dotted pink line and the dotted blue line, for a given level on the real house price index (Y axis). Eyeballing the chart, you can see there are something like 30,000-40,000 “missing” homes, every single year, were 2010-18 vintage Ireland to build with the same enthusiasm as 1995-2007 vintage Ireland. 

One of the main reasons is certain to be higher construction costs. High construction costs have the effect of dampening supply for any given level of house prices. And Ronan Lyons found construction costs rose between 70-90 per cent between 2007 and 2020.

So to solve the housing affordability crisis we need more houses, and to get more houses we need construction costs to go down. 

Why have construction costs gone up? This is a big and difficult question. Higher standards and tighter regulations are likely part of the picture.

But construction costs have risen all over the world in recent years, even in places where regulations haven’t gotten tighter. The biggest contributor to rising construction costs is almost certainly the construction industry’s stagnant productivity, i.e. output per hour worked.

Construction is an industry that has stubbornly refused to get more productive. That’s true in Ireland and all over the rich world. Where other industries are investing in robots and AI and the like, construction companies are building houses in basically the exact same way as 30 years ago. 

Why haven’t construction companies become much more productive? In other words, why haven’t they figured out ways to automate jobs on the site? 

One answer is that construction sites are by their nature dynamic; no two of of them are the same. So it’s hard to come up with standardised processes. Another might be that construction is quite highly regulated. Materials, safety and other standards change a lot, that adds another layer of complexity. Then there’s the fact that even the biggest contractors operate incredibly lean, on 2-3 per cent margins. That doesn’t leave much room for an R&D budget.

Architect Justin Kinsella. Pic. Bryan Meade

Justin Kinsella and Paul Kelly have worked as architects in the British and Irish construction industry for decades. The pandemic gave them a bit of time and space to think about trying something different. 

They had heard of COBOD, a Danish firm working on a 3D printed concrete technology. The technology could, in theory, be used to replace blocks, bricks and cladding in what’s called the superstructure of a house. It had the potential to build homes quicker and cheaper than the existing technology, which is more or less unchanged for decades. 

So in early 2021 they flew out to Denmark, and started a conversation about bringing COBOD’s technology to Ireland. In the following months, they secured exclusive rights to the COBOD kit for the UK and Ireland. The company they established is called HTL.tech.

“A lot of other adopters of this technology are coming from a technology background, and not a construction background. And I think that’s where we add value. We’re looking at the reality of building in Ireland,” said Kinsella.

Since then, HTL has been in stealth mode, in an unmarked industrial unit near Drogheda, where they’ve been tinkering with the machine and adapting it to fit with UK and Irish building regulations. Now, having made their first sale, they’re eager to share what they’ve been working on.

How the rig works

The walls and roof of a building are called its superstructure. And the superstructure is a particularly important part of any project, for two reasons. 

First, as we’ll see, it makes up a big part of the overall cost of building a house. Secondly, most other elements of the construction process are downstream of it. By making the superstructure neater, more predictable, and quicker, the rest of the construction process is made more efficient too. 

The elements of a house are as follows: the substructure (i.e. the foundation), the superstructure (i.e. the walls floor and roof), the completion (windows, doors, skirtings etc), the finishes (painted walls, tiled bathrooms etc), the fittings (kitchen, wardrobes etc), and the services (plumbing, electrics etc). Throw in the costs associated with getting the site ready — like demolition drainage and groundwork — and according to the Society of Chartered Surveyors of Ireland, in 2020 it cost €179,000 to build a standard house. And of the €179,000, SCSI said the superstructure cost €56,000. That €56,000 outlay is the problem HTL was set up to solve. 

The machine is made up of five metre long metal beams. They arrive on site on the back of a lorry, are bolted together into a frame and secured to the foundation. The thing takes about a day and a half to set up. 

The HTL house printing machine. Pic: Bryan Meade

When finished, the frame takes the form of four vertical posts, one at each corner of the structure. Along each end, there’s a horizontal “cross bar”, linking the vertical posts. And there’s another horizontal bar that links the two cross bars, looking something like a ceiling joist. That’s where the 3D printer nozzle fits. The beams are designed to slide smoothly across each other, so that the nozzle can be precisely positioned over any point on the superstructure. 

At its snout, the 3D printer looks like a finely machined piece of lab equipment. At the point where it connects to the “joist”, it has a bulky metal piece, which connects by a hose to the cement outside. The cement is processed by a special tool, about the size of a coffee cart, so that it’s at just the right consistency to flow through the machine – and set – at the right rate.

The machine builds up the concrete walls of the house layer by concrete layer – like the layers of a tree. Each layer is 50mm thick. Over the course or a day, on a typical house, it can build a three metre wall. 

Some 11 days after the machine arrives on site, HTL say you’ll have a completed superstructure. It’s operated by four or five people, and it does the work of eight to ten people. That’s more than 50 per cent faster than the traditional route, with blocks and bricks. And, HTL said, it’s about 25 per cent cheaper. On a standard house, that’s a saving of €14,000.

HTL House printing machine . Pic. Bryan Meade 15/02/2022

When pressed, HTL said the 25 per cent saving should just be the start. That’s because the machine makes building more standardised and predictable. All the other workers can rely on the superstructure being completed precisely, and quickly, and that helps their work.

“You can more confidently sequence the building and the construction on site,” said Kinsella. “You can have materials arriving on site, and know they’ll be there at the exact stage when they’ll be fitted. You don’t have to store massive quantities of gear on site.

“You’re saving the time of sending someone out to measure openings for windows, because you can standardise and order everything up front. The same on flooring — you can get high-performance off site flooring, and bring them to site in the most efficient way.

“When you get into the scale of housing developments in the UK, and start factoring the multiples in speed of delivery, it starts to make a lot of sense.”

And leaving aside efficiency, Kinsella said HTL’s rig is better for the environment. The resultant homes are better insulated, and much less cement is wasted during construction. Kinsella said there’s a 70 per cent reduction in cement waste, which is important because the manufacture of cement is a particularly dirty industry. 

“Everything from architects, engineers, M&E, everyone down to windows and floors, is designed in 3D,” said Kelly. “But the biggest component – the superstructure – is currently built by hand. That creates issue with time and accuracy. Whereas [the 3D printed version] can’t deviate from the 3D model.”

Paul Kelly. Pic. Bryan Meade

Right now, HTL’s goal is just to show that it can replicate existing outputs more efficiently. But the COBOD machine can do more than build the equivalent of a block wall, more cheaply.

To comply with building regulations that were designed for concrete blocks, HTL’s version of a wall will end up being more than five times stronger than a concrete wall. That creates scope, in time, to make the walls much thinner and thus, save more money. 

Then there’s the potential for exciting new designs and forms. The rig can create curves as easily as straight lines. That frees designers to try new forms and new designs. It would allow them to build a terrace of unique houses for almost the same cost as making each of them identical. 

Another feature of the rig is that it’s much easier to get a cement truck to a remote location than alternative technologies like blocks or pre-cast concrete. This enables new types of construction that would not have been viable before. For example, in a remote part of Denmark, General Electric used a COBOD rig to build a giant 30m concrete podium to house a wind turbine. It would have been much more difficult to truck pre-cast concrete out there.

*****

Ireland’s cost disease

Ireland is afflicted with something called cost disease. First described by William Baumol in 1967, cost disease refers to the way labour-intensive industries are doomed to have high and growing costs as the economy gets richer.

The way it works is that there are two types of industries: progressive ones and stagnant ones. The progressive ones are industries that are getting more productive every year – like manufacturing. They make the economy richer. The stagnant ones are the ones whose productivity doesn’t improve from year to year. Examples would include education and construction.

Progressive industries are the ones that can be automated. Stagnant industries are the ones that required a lot of workers. 

This is how high costs infect stagnant industries: we still need teachers and builders, whether or not their productivity is increasing every year. They have to be paid the going rate. And the going rate is set by the wages builders and teachers could have made if they worked in the progressive industries. 

So the cost of builders go up every year, even though their productivity does not. The result of this is that, over time, spending on the stagnant sectors grows to take a bigger and bigger share of the economy. 

This is bad because it prevents us from enjoying the fruits of the wealth created by the productive sectors. Because of cost disease, much of the surplus created by productive industries flows to unproductive ones, leaving consumers no better off. 

This is not just an Irish problem. All across the rich world the share of the economy going to labour-intensive industries, i.e. ones whose productivity doesn’t tend to rise much, is getting bigger and bigger. Their costs are getting higher and higher, without much commensurate improvement in quality. Look no further than Ireland’s health service for an example.

The only way to cure cost disease is to figure out a way to automate some tasks. As long as they’re labour intensive, they’ll continue to get more and more expensive every year. 

Market and model

HTL didn’t design the rig from scratch. It was designed by COBOD in Copenhagen. COBOD’s strategy is to licence out its technology to the likes of HTL, because construction is a heavily regulated industry and there’s a fair bit of work to be done making the rig compliant and functional under local building codes. HTL has attained the exclusive rights to the COBOD rig in UK and Ireland. 

3D printed concrete, though still in its infancy, is starting to be used around the world. An American company called ICON is one of the biggest players. It works with the US Department of Defence, NASA, as well as on ordinary residential homes

HTL’s initial target market is the large homebuilders in the UK and Ireland, companies that build thousands of homes each year. What’s interesting though is that the rigs are small and neat. Workers can be trained on them in six or seven weeks.

Brian McDonnell working on the HTL rig. PIC: Bryan Meade

This means small and mid sized builders, as well as big ones, could benefit. Modular home factories, by contrast, are much more capital intensive. 3D printing might be accessible to small developers in a way modular homes will never be.

It’s easy to make a case for a limited application of the machines: slightly improving the productivity of the largest house builders. But, in time, it’s possible to see it being central to the workflow of all kinds of builders, big and small, resulting in cheaper homes and buildings for customers.