Ian Mulheirn recently published a series of three well-researched and well-referenced blogs on the UK housing market. They challenge some widely-held views.
In this blog I will focus on his first blog post on surplus homes.
My post is too long and it’s not even complete. As I solve problems it will get shorter and more complete. I've since published a second post arguing that "ONS household count estimates are either wrong or useless. They may be both".
It’s taken me weeks to understand, then investigate, and then reproduce what he’s found. I’m left with more questions than answers. I’m sharing my work in the hope that others can both help to improve on it and build on it.
The problem is simple.
I think that most parts of England & Wales (from now on England, sorry) have built too few homes in recent decades. Ian’s first blog says that England has been building more than enough homes.
I think that most parts of England are currently planning too few homes. Ian’s first blog says that England is currently planning more than enough.
There's a lot that we agree on. I have long said that there is no UK housing crisis. Rather there are different challenges in different parts of the country that require different solutions. But I still think Ian's case needs investigating. So let’s investigate!
We don’t know how many households there are in England. But we do have good estimates.
Two national bodies estimate the number of households in England; the DCLG and the ONS. They use different methods that give different answers and they report their estimates for different time periods and geographies. The two estimates are most similar in 2001 and 2011 when censuses greatly increase the amount of data that they can work with.
DCLG data provides both future projections of household number and data at local authority level. The ONS provide no future projections and only give a single number for the UK. From this I calculate a number for England by assuming that 83.9% of UK households live there.
Let’s compare the data.
Good news. The historical trends are the same. In 2001 and 2011, aided by the census, both estimates are the same. And I can reproduce Ian’s result showing that ONS and DCLG household estimates diverge after 2011.
The divergence is crucial to Ian’s point and needs looking at in more details.
The DCLG household estimate is much more useful in trying to understand the housing market because it provides data at local authority level and because it makes projections into the future.
But Ian says that the ONS household estimate is more accurate and that’s why he uses it and not DCLG household estimates from 2011 on.
There are dozens of long pdfs comparing the approaches. I am really slow at understanding these type of things and it will take me a lot of time. To make things even more complicated the ONS changed how it defines a household in 2011 — exactly when all these troubles start. If I need to read all the reports I will, but I think it’s useful to share my concerns now.
I’m worried about the ONS estimates because they deviate so strongly from the historical trend up to 2011. In the five years to 2010 the number of UK households increased by 4.1%. In the five years since 2011 the ONS estimates that the number of UK households has increased by just 2.6%.
One possible explanation for this would be a large decrease in UK household formation due to lower net immigration. And yet net immigration has been the same (1.2m) since 2011 as it was in the five years before.
There are other possible explanations. The most likely is hinted at in the huge increase in multi-family households and the decrease in lone parent households since 2011. A reasonable hypothesis is that that more families chose to live with other families and that more unhappy couples chose not to separate because housing is so expensive. It would be possible to test that hypothesis if the ONS produced estimates by local authority. But they do not.
Even if we assume that the ONS estimate is correct it is not sensible to scale down DCLG local authority estimates by a constant so that the total for England matches the ONS estimate. High housing costs in London that encourage young professionals to live together do not have the same impact in Bradford where housing remains cheap.
Last of all there is the possibility that the ONS estimates are further from the truth than the DCLG estimates. Has our population really grown at a higher than normal rate and yet the number of households has grown at a lower than normal rate? It doesn’t feel right to me. And later I’ll present more data that suggests that the ONS data is weak.
For the rest of this piece I will use DCLG household estimates . Between 1996 and 2011 it is uncontested. Beyond that, it is the least worst data that we have on a useful scale: the local authority.
Let’s move on to housing stock.
We don’t know how many homes there are in England. But we do have good estimates. They are provided by DCLG in their table 125.
Calculating the number of surplus homes is easy. Take the number homes and subtract the number of households. Doing this I reproduce Ian’s graph up until 2011. But then where he shows a sharp increase in surplus dwellings I see a continued decline.
At this point let me introduce a new time series: vacant homes. The DCLG report two measures of vacant homes — vacant, and long-term vacant. A vacant home becomes long-term vacant after six months. Both measures are estimates. In Leeds this data is collected by the local government using their own methods, but I don’t know how it is collected elsewhere.
Until 2013, data on long-term vacant homes was very likely to be accurate. Council tax exemptions or reductions were given on vacant homes. This providied an incentive for owners to inform the local government that homes were empty. In 2013 the law changed and local governments could charge full or higher council tax on empty homes. This pushed property owners to fill empty homes more quickly. But it also gave them an incentive to lie or mislead.
It is obvious in the graph that the number of vacant homes fell slightly more than the trend in the year to 2013. But the effect is small. The similar pattern in the vacant homes measure and the surplus homes measure seem to support the DCLG household estimate above the ONS estimate. Certainly we do not see any increase in empty homes in 2012 where the ONS household estimates suggest that the number of surplus homes rose sharply.
I have been screaming for nearly five years that England does not have a housing market. The housing challenges facing Sunderland are different from those in Manchester and both are different from those facing London.
For Sunderland the data is clear. The number of surplus dwellings has not fallen, and nor has the number of vacant homes.
Manchester is completely different. The number of surplus dwellings has fallen and is now close to zero. And really that's all that matters. Because people like me aren't all that bothered about increasing the number of homes built in Sunderland; they've done a good job. But I would like to build more in Manchester, and to keep building enough in Leeds. I don't want a housing crisis like London has where any increase in productivity is swallowed up by higher rents and given to undeserving landowners.
This blog post is already extremely long and it's not even half of what I've found in this search. As I said at the start, it will get shorter and more complete as I find answers to these problems. But in advance here a few more things I've been looking at.
ONS household estimates start to diverge from DCLG estimates 2011, so it’s interesting that the ONS definition of a household changed in that year.
The ONS notes says that “for 1996 to 2010, a household is defined as a person living alone, or a group of people living at the same address who have the address as their only or main residence and either share one main meal a day or share living accommodation (or both)” and “for 2011 onwards it is defined as one person living alone, or a group of people (not necessarily related) living at the same address who share cooking facilities and share a living room or sitting room or dining area”.
I am not sure if the DCLG definition of a household also changed in 2011.
I calculate how many surplus homes there are by subtracting the number of dwellings from the number of households. But there are complexities that this doesn’t take into account. International students do not seem to be counted as households. And blocks of student flats do not always seems to be counted as dwellings. The status of second homes, both rented and owned, is also confusing.
Leeds collects and releases excellent data on empty homes by ward. And looking at these graphs brings up even more problems.