A great post. I'll definitely be referring people to this in the future.
Housing remains a problem were the evidence points to a clear solution and the commentariat searches for a more complex one.
We really need a shorthand to refer to these two types of landbanking: landbanking as a way of creating a planning pipeline in response to unreliable planning permission and landbanking as a form of pure speculation.
In theory, speculation should exist with any finite good that does not become obsolete or decay. All things being equal, the economy of the future has more money than the economy of today. So in a truly competitive market, I would be surprised not to see some speculative land banking, given how rare a good with the qualities of land is in the economy.
We do see plenty of speculative land hoarding like behaviours in video games were these conditions of competition and non-decay do exist.
However, the UK is far from a truly competitive market. UK planning is a form of alchemy, transmuting land from a good that does not become obsolete or decay to a good that changes in value year-on-year depending on the whims of governments and officials.
Planning remains the key limit in medium to longterm horizons for housebuilding. Pipeline landbanking is a rational response. In 'Home Truths' Liam Hannigan talks about in detail how the UK has ended up with an usually large number of big housing firms compared to peer nations because scale is necessary to support a planning pipeline. If the UK adopted Japanese style planning permission lots of small building firms would be able to take on the risk of more projects. The pipeline landbanking would mostly disappear.
The UK probably never gets anywhere near the limit where purely speculative landbanking makes sense.
I don’t quite get the distinction you make for the two reasons to delay. Delay is just delay at all points in time.
I think you might also be missing a key bit of economics - the rate of return on waiting has to be the same as the rate of return from building in a market equilibrium. This will imply many under-utilised properties that are currently feasible for development being delayed.
Perhaps there are a few more subtleties, e.g for point 3) and subsequent points, the expected price growth doesn’t have to be positive, just more positive than pushing out extra homes now. For example, when prices are falling in a thin market with few buyers, pushing out new homes faster will have a larger price effect than in thicker rising markets. Avoiding that price effect is also part of the gain from waiting.
On 4) and in general, the holding costs don’t have to be low, as the equilibrium pricing outcome in the land/site market is that the net return to land not developed is similar in risk-adjusted terms to other investments. The price will embed holding costs such that money spent on land to wait will get roughly the same risk-adjusted return.
On 5) maybe some building approvals are left unbuilt, many planning approvals are, and an enormous number of sites are held waiting for future planning permissions, even though starting development now would be higher value than the current use.
Does that make sense? Most cities have enormous numbers of homes that can be built under their planning schemes, and many also have huge stocks of planning-approved projects (if planning approvals are relatively low cost and can be sought early and modified later, etc)
I’m just not sure whether “deliberately delay” and “put in a future pipeline” are different. I mean, when is this future pipeline? How do we know it won’t be delayed too? Isn’t building it at an unspecified future time instead of immediately also deliberate?
1. It's a little stronger than 'deliberate delay' - it's 'delaying specifically because it makes you more profit than building would' (as Michael suggests, the two would imply differing exposure to antitrust action)
2. Development pipelines are very much not unspecified with respect to build-out times - they are mapped out quite explicitly (T+5mo we begin work on this site, T+20mo we begin work on this site...)
If putting a site in a pipeline made less profit than building now you wouldn’t do it, no? You’d just build everything as quick as possible then wait until the next project.
"...to the extent that the government itself says that these numbers should not be used to estimate the number of unbuilt dwellings with permission" -- nowhere in that link is that said. In fact it says they're approved statistics with "low risk" of being low quality!
That's in the section "Data on housing approvals provided by Glenigan", who were used for data briefly early in the pandemic in 2020. It doesn't apply to the rest of the data.
The rest of the data does not contain unit-level housing approval data - Glenigan is used as the source for that. You are also wrong that 'nowhere in that link is that said'. You are correct that Glenigan data is separate from the main planning applications dataset – but that's not the point I was making.
You seem confused over what land banking is. It's exactly what Letwin described – not building homes at a rate which would cause local prices to fall. Of course private house builders won't do that; that's why we used to have social housing construction, which specifically did.
Your insights are both thought-provoking and clearly conveyed.
Looking forward to reading more of your work!
Great post! It could only be better if you linked to my Substack. 😉
A great post. I'll definitely be referring people to this in the future.
Housing remains a problem were the evidence points to a clear solution and the commentariat searches for a more complex one.
We really need a shorthand to refer to these two types of landbanking: landbanking as a way of creating a planning pipeline in response to unreliable planning permission and landbanking as a form of pure speculation.
In theory, speculation should exist with any finite good that does not become obsolete or decay. All things being equal, the economy of the future has more money than the economy of today. So in a truly competitive market, I would be surprised not to see some speculative land banking, given how rare a good with the qualities of land is in the economy.
We do see plenty of speculative land hoarding like behaviours in video games were these conditions of competition and non-decay do exist.
(See "Land value tax in online games and virtual worlds: A how-to guide" https://www.gamedeveloper.com/design/land-value-tax-in-online-games-and-virtual-worlds-a-how-to-guide)
However, the UK is far from a truly competitive market. UK planning is a form of alchemy, transmuting land from a good that does not become obsolete or decay to a good that changes in value year-on-year depending on the whims of governments and officials.
Planning remains the key limit in medium to longterm horizons for housebuilding. Pipeline landbanking is a rational response. In 'Home Truths' Liam Hannigan talks about in detail how the UK has ended up with an usually large number of big housing firms compared to peer nations because scale is necessary to support a planning pipeline. If the UK adopted Japanese style planning permission lots of small building firms would be able to take on the risk of more projects. The pipeline landbanking would mostly disappear.
The UK probably never gets anywhere near the limit where purely speculative landbanking makes sense.
great post. this will be more go-to link for refuting this bs! thanks
Very thought provoking.
I don’t quite get the distinction you make for the two reasons to delay. Delay is just delay at all points in time.
I think you might also be missing a key bit of economics - the rate of return on waiting has to be the same as the rate of return from building in a market equilibrium. This will imply many under-utilised properties that are currently feasible for development being delayed.
I explain here
https://www.fresheconomicthinking.com/p/explainer-markets-efficiently-delay
I’d be interested in your comments/response to this article.
Just so I understand your argument:
(1) At equilibrium, the marginal developer will be indifferent between building now and waiting
(2) Therefore, the NPV of immediate development must equal the expected NPV of deferring if a house is going to be built
(3) If expected price growth is positive and the costs of holding are low, expected NPV of deferring > NPV of immediate development
(4) At equilibrium, there are cases where expected price growth is positive and the costs of holding are low
(5) Therefore, at equilibrium, there are unused permissions the building of which is being deferred
Is this correct?
A good summary.
Perhaps there are a few more subtleties, e.g for point 3) and subsequent points, the expected price growth doesn’t have to be positive, just more positive than pushing out extra homes now. For example, when prices are falling in a thin market with few buyers, pushing out new homes faster will have a larger price effect than in thicker rising markets. Avoiding that price effect is also part of the gain from waiting.
On 4) and in general, the holding costs don’t have to be low, as the equilibrium pricing outcome in the land/site market is that the net return to land not developed is similar in risk-adjusted terms to other investments. The price will embed holding costs such that money spent on land to wait will get roughly the same risk-adjusted return.
On 5) maybe some building approvals are left unbuilt, many planning approvals are, and an enormous number of sites are held waiting for future planning permissions, even though starting development now would be higher value than the current use.
Does that make sense? Most cities have enormous numbers of homes that can be built under their planning schemes, and many also have huge stocks of planning-approved projects (if planning approvals are relatively low cost and can be sought early and modified later, etc)
Isn't the distinction just about antitrust implications?
I’m just not sure whether “deliberately delay” and “put in a future pipeline” are different. I mean, when is this future pipeline? How do we know it won’t be delayed too? Isn’t building it at an unspecified future time instead of immediately also deliberate?
Two points:
1. It's a little stronger than 'deliberate delay' - it's 'delaying specifically because it makes you more profit than building would' (as Michael suggests, the two would imply differing exposure to antitrust action)
2. Development pipelines are very much not unspecified with respect to build-out times - they are mapped out quite explicitly (T+5mo we begin work on this site, T+20mo we begin work on this site...)
If putting a site in a pipeline made less profit than building now you wouldn’t do it, no? You’d just build everything as quick as possible then wait until the next project.
It's a point about market power, ie. the magnitude of delay varies with the competitiveness of the market.
"...to the extent that the government itself says that these numbers should not be used to estimate the number of unbuilt dwellings with permission" -- nowhere in that link is that said. In fact it says they're approved statistics with "low risk" of being low quality!
Section 4, "It is not possible to use these annual figures to estimate the total stock of units on uncompleted sites with permission."
But thank you for being such a good-faith interlocutor.
That's in the section "Data on housing approvals provided by Glenigan", who were used for data briefly early in the pandemic in 2020. It doesn't apply to the rest of the data.
The rest of the data does not contain unit-level housing approval data - Glenigan is used as the source for that. You are also wrong that 'nowhere in that link is that said'. You are correct that Glenigan data is separate from the main planning applications dataset – but that's not the point I was making.
You seem confused over what land banking is. It's exactly what Letwin described – not building homes at a rate which would cause local prices to fall. Of course private house builders won't do that; that's why we used to have social housing construction, which specifically did.