Monday, 24 September 2018

There is a housing crisis: there are not enough houses being built and they are becoming ever less affordable; and they are getting smaller…

[By Prof Paul Cheshire]

As a species we are very well honed not to face unpleasant facts. We mock the ostrich but humans have head-burying down to an art. Worried you might have cancer? Don’t worry the doctor – she might confirm it.

But we can do better than deny the danger: we can create for ourselves an alternative story to reassure. Climate change is a major threat? Roll out Nigel Lawson to paint an alternative reality. Brexit is likely to be seriously damaging to our prosperity? Roll out Boris Johnson to imagine a rosier world: oh – or, of course, the all-purpose danger denier – Nigel Lawson.

And so it is with housing. A long term and endemic lack of housing supply? Ian Mulheirn will explain why that is not so. And now houses are getting too small and there isn’t room to swing a cat? No worries: Dr Chris Foye, from the newly funded Collaborative Centre for Housing Evidence can resolve your worries. Contrary to what other researchers and agencies have been saying he can reassure you not only are houses really getting bigger but we do not want more space in houses anyway. So that is both ends covered: we do not have to worry about shrinking living space.

We dealt with Ian Mulheirn’s ‘no shortage’ claims in a recent blog. So what about this most recent claim? All the authorities and surveys of which we are aware show significant falls in the size of new houses built in Britain (or England depending on source); and at the same time they show how much smaller new build houses are in Britain/England than in other comparable countries: 38% smaller than in Germany and 40% smaller than in the land-strapped Netherlands in 2005.

It is true that there is a wealth of folk legend in measuring house sizes and some factoids of dubious provenance (here's an example of how a number of 76 metre squared for the average size of a UK house or new build came to be built in to ‘knowledge’.) We may legitimately have concerns as to the reliability of comparative house size data but the pattern does seem to be consistent. The most recent official publication giving EU wide comparisons shows English houses were small by rich EU country standards and getting smaller: the average size of existing houses was given as 87m2 while that of more recently built houses was 83m2 with only Romania and Italy having smaller new build houses; and Denmark having new homes averaging 132m2.

Dr Foye however, claims new build houses are getting bigger in England – from 88m2 in 2004 to 90m2 in 2016. Examining the details of the source he cites for this however reveals it is spurious. It relates to the size of homes sold: very different from new homes. As the ONS helpfully explain:
… floor space [has] seen small increases over the period with the biggest shift seen between 2008 and 2009. During this period there was an increase in the proportion of detached properties purchased … and a respective fall in the proportion of flats. As flats tend to be smaller than houses this contributed to the growth seen in floor space … between 2008 and 2009. Since 2012, while the proportion of detached properties has remained broadly consistent, the proportion of flats purchased has increased. This has reduced the average … floor space slightly, but it is still above 2004 levels.

In other words, what Dr Foye is talking about is not the size of new build houses but the size of houses currently being sold. And that will, of course, be strongly influenced by the composition of sales. The claimed increase in new house sizes is an artefact of more detached properties being sold during the financial crisis years; not evidence that house sizes are increasing.
A longer term but less well documented source is here. This is not peer reviewed research but was based on samples of houses available on popular property websites and using their dates of construction and details to derive average size of rooms and total floor area by decade built. This, too, could be subject to composition bias but it is less likely to have been a problem with this measure since that would require there to have been systematic differences in the composition of sales varying by year of construction. The conclusion was that houses were biggest if built in the 1970s when they averaged 83m2. Floor area then fell pretty steadily decade by decade to post 2010 when the size was just 68m2 – smaller than those built in the 1930s.

One possible reason house sizes could be getting smaller is that family sizes are falling. But at the same time over the 50 years since 1970 incomes have risen and we know there is a strong income elasticity of demand for space implying house sizes should have increased. And they did in almost all other countries with rising real incomes over time. So falling real incomes since the financial crisis might explain a post-2010 size reduction in new build sizes but over the longer term the expectation would be – if supply was able to respond to the changing structure of demand – new build houses would have got a lot bigger between 1970 and 2015. Indeed there is even some evidence that, independently of income, the demand for space in houses increases with age. And as we know the population of England has been aging since 1970 – this would again suggest that if supply was not constrained in some relevant way, houses would have got bigger.

To be fair to Dr Foye his claim that we really do not want bigger houses anyway seems to be more a subeditor pitching for a good headline than it reflects the contents of his research. There may be an argument to be had - what does one mean by ‘want’? But it is my judgement that at least in terms of effective demand, the evidence is overwhelmingly that as people get richer they demand more space in homes and that the UK planning system frustrates this demand by constraining land supply.


Chris Foye said...

Thanks, Paul. I need to split this comment into two parts to get it through the system.

PART ONE: Let me clarify a couple of things before discussing that stat. First, this article was based mostly on my own PhD research, and does not reflect the views of the Collaborative Centre for Housing Evidence. Second, the primary focus of the article is on the relationship between size of living space and subjective well-being. Neo-classical economics tells us this relationship is very simple: individuals have certain preferences – preference for a garden, or detached home, or more living space- which when met, leads to an increase in happiness. The objective of policy should therefore be to satisfy these preferences insofar as possible – through maximising incomes, increasing the supply of housing – thus producing the greatest happiness for the greatest number. In the BBC article, which was written for a general audience, I seek to show that the relationship is not that simple.

First, Individuals (partly) adapt to increases and improvements in living space, meaning that a substantial portion of the uplift in happiness associated with having a bigger house is likely to be fleeting. On this point, see and

Second, individuals not only care about their absolute house size but also their relative house size. In short, house size is a positional good. To illustrate, let us imagine that over the next decade, the most space-rich ten percent of the UK population gained an extra room while the rest of the population’s living space stayed flat. All other things constant, neo-classical economics says this would increase societal well-being. I’m not so sure. Analysis from the USA suggests that this could harm societal well-being because the other 90% of the population would feel worse off. See

I’m not the first to make this argument: the economist Robert Frank was making it a decade ago. He suggests that we would all be happier if we spent a higher proportion of our national income on non-positional goods like longer vacations and shorter commutes which aren’t associated with the same negative externalities. See book review here;

More research is needed to understand the positional nature of housing in the UK, and I am doing some work to address this. For example, we don’t know whom people in the UK compare their house size with: their neighbours? their friends? the super-rich?. Nor do we know how this positional good effect operates: do higher levels of living space lead to more home-centred social norms to which the space-poor cannot conform? Or do people just feel worse off after seeing others with bigger houses than themselves?

Chris Foye said...

PART TWO As I say in the article, there is an overwhelming case for providing more genuinely affordable housing for those suffering the most cramped, unaffordable living conditions. This group are suffering too much to wait for academics to conduct further research into the positional nature of housing. But taking a long-term perspective, I’m just not so sure that continuous increases in the size of living space will lead to continuous improvements in societal well-being, which is why I say that relationship is open for debate. Of course, there are good arguments for providing more housing regardless of the relationship between size of living space and subjective well-being which you know more about than me. For example, as John Myers and others frequently argue, allowing London to densify would boost the productivity of the economy which would allow us to spend more money on the NHS, vacation time, welfare – indisputably good things. I think there are good arguments for increasingly supply in those areas with the greatest affordability problems but we should think twice before assuming that continuous increases in the size of living space will lead to continuous improvements in societal well-being.

Now, on that stat. Your blogpost focusses on one particular statement I make, "But the reality is that living spaces in England and Wales are actually larger than ever, with the average home increasing from 88 to 90 square metres between 2004 and 2016." I think you are correct to criticise this statement, as it conflates two things: the average size of the total housing stock and the average size of new housing. So please let me distinguish between the two.

In terms of total housing stock, which I am most interested in, all the evidence I have seen suggests that average living space in the UK has increased over time. Using Census data, Tunstall (2015) showed that between 1911-2011, the population grew by half, but the number of rooms tripled. Using EHS, Ian Mulheirn's analysis also indicates that “mean floor space per person per dwelling in England has increased over the last 20 years, by 2 to 3%, to 38.5 square metres per person per dwelling” Thus I would stand by the statement that “living spaces in England and Wales are actually larger than ever”, and would be interested in seeing any evidence that suggests otherwise. Tunstall's link can be found here, and Ian Mulheirn's piece is cited by you in the blog.

In terms of the new stock, which your article focusses on, I accept your point that the ONS figure is not ideal. It was not my first choice and I will avoid it in future. That said, Figure 4 in this piece by Neal Hudson suggests homes built post-2002 are slightly bigger than older stock (but more unequally distributed) -

And Table 2 in this report from MHCLG shows that new build homes in England and Wales are on average larger than the existing stock:

However, I don’t think debating whether new homes are getting bigger or smaller is a particularly good use of time, because it ignores changes in household structure. If the average new home is getting smaller because more people want to live on their own, then I don’t think this is necessarily cause for concern. Surely what matters is average living spaces per person which are “larger than ever”.