19 May 2015

More evidence on why high house prices are driven by credit availability, not a shortage of houses


In a recent CapX article, I argued that there is no UK “housing crisis” and there never was one, if by “housing crisis” one means that the country has run out of houses or that the cost of living in a house (e.g. the cost of renting) has risen enormously.  I claimed that house prices rose as economic conditions seemed strong and mortgage credit was easily available and that house prices fell back when economic conditions were poorer (when there was a recession in 2008-09 and subsequent poor growth) and mortgage credit dried up.  I explained that the reason it was originally believed there was a housing crisis was because official population data in the late 1990s seriously over-estimated the population (by 900,000 persons).

My article has produced predictable outrage.  I want to respond to some of the points made, here, and to offer some further evidence to support my key conclusions.

A first point that some commentators raise is that housebuilding volumes have fallen significantly in recent years.  See Figure 1.

Figure 1: Housing completions – All Dwellings


Source: https://www.gov.uk/government/statistical-data-sets/live-tables-on-house-building, Table 211

We can see here that housebuilding volumes had risen to a nearly 30 year high in 2007 of about 230,000 dwellings, but then as house prices crashed and the economy went into recession, volumes fell to below 150,000 since 2010.  Construction volumes quite often fall in recessions and house price crashes — we see that for example from the late 1980s in the figures above.  And it’s certainly arguable that construction volumes in the UK have not been as responsive as they might have been to changes in house prices.  Indeed, my Europe Economics colleagues wrote a report last year on policy reforms that might improve the responsiveness of construction.

But the fact that construction volumes have been fairly low since 2007 manifestly does not mean we had a housing crisis from the late 1990s (which was everyone started saying we had one).  Furthermore, the fact that volumes are relatively low now does not mean we’ve run out now.  If we have surplus houses now (which we do, and always have done, at least since the late 1980s) then under-building now might mean eating into that surplus.  The most it might mean is that we might eventually run out if housebuilding doesn’t pick up.

A second point some commentators have raised is that the tale of additional housebuilding keeping pace (or more than keeping pace) with the rise in the number of households seems to be in tension with the rhetoric of the immigration debate.  Aren’t we continuously told that the country is full and that house prices are high because of all these immigrants coming in?  Or if not immigrants, then at least rising population because of increasing longevity?

Well, here are some facts.  The UK population in 1981 was 56.3 million.  By 2011 that had risen to 63.7 million.  That means that average annual growth in the population was 247,000 people.  The average household size over that thirty year period was about 2.4 persons.  So the average number of additional houses required to accommodate the rise in the population (setting aside any changes in the nature of households such as divorce or kids living with their parents into their 20s) would have been 102,800 per year.  The actual number of additional new dwellings built per year from 1981 to 2011 was 192,300 — nearly twice as many as required to accommodate population growth.  That isn’t quite the whole story, since there would have been some demolitions and some existing single dwellings converted into multiple dwellings or vice versa, but it does provide the core picture: over those thirty years there were vastly more new houses built than were required by any population rise.

Those that currently argue that we do not have enough houses to cover all the population claim some other things about how the number of households should be counted and how new household formation has changed over time.  I’ll have some things to say about that in a moment.  But before I do so, I want to re-emphasize a point that I made in my previous article.  When the notion that there was a housing crisis first got going, it did so on the basis of the very same statistics I quoted in my previous article about the ratio of houses to households.  At that point, no-one felt they needed to argue about “suppressed household formation” or anything of that sort.  The “housing crisis” was something quite specific and concrete – there were X houses in London and the South-East of England, and Y households, and Y was a lot more than X.  And those statistics were wrong.

It turned out, when we did the 2001 Census, that the population of England and Wales was about 900,000 fewer people than we’d thought.  That utterly transformed the picture on the ratio of houses to households, and should have totally transformed the public debate.  Those that wanted to argue we had a “housing crisis” were of course entitled to do so.  But they should not have been able to pretend that nothing was changed by it turning out that there were nearly a million fewer people in the country than we’d thought there were.  If we still had a housing crisis even with nearly a million fewer people, imagine what sort of crisis we’d have had if the population had, in fact, been nearly a million higher?

Now the fact that we didn’t have the housing crisis that was alleged in the late 1990s — the crisis of having fewer houses than households as per the data at that time; and, I repeat, we indisputably did not have that crisis, even if you want to claim we had some other one, because that crisis was based on demonstrably wrong data — does not prove that we do not have some other housing crisis.  There are those that claim the Census data dramatically under-states the number of households because household formation is suppressed by high house prices.  (But note once again: this difference between the official estimates of the households done in 2000 and the actual households was not some matter of how households are defined.  It was because the population of individuals was nearly one million lower than thought.)

My first, minor, comment on this claim that I don’t really see why household break-up driven by divorce counts as a “real” increases in the number of households whilst household fusion via kids living with their parents doesn’t count as a “real” non-increase.  If Mum and Dad get divorced and Dad lives elsewhere whilst Son lives with Mum until he’s 25, I don’t see why it’s any more correct to say that Mum and Son are two “real” households than to say that Mum and Dad are “really” two half-households.

My second remark would be that the original “housing crisis” was something concrete – there were fewer households than houses.  This “housing crisis” is something much less tangible – something to do with households that are alleged never to have formed but ought, in some sense, to have done.  And it’s not that the average household size has been rising – it’s actually fallen very marginally over the past 30 years (from a little over 2.4 to around 2.3).  The thought is that it “ought” to have fallen further.  And these extra, smaller, households have not been deterred from forming by housing costs in terms of rents being high – as we saw in my previous article, rental prices have not risen especially rapidly for many years.  Rather, the idea is that 25 year old Son doesn’t think it’s worth moving out from Mum and Dad’s just to rent elsewhere but would have chosen to move out if he’d been able to afford a house.

Next, the relationship between high house prices and stay-at-home 20-34 year olds is sketchy.  Here is the relevant ONS chart.


Figure 2: Young adults aged 20-34 living with parents in the UK, 1996-2013

We can see here that by the peak of house prices, in 2007, stay-with-parents 20-34 year-olds were only about as numerous as they were in 1996.  The number of them rose dramatically after 2007, as house prices fell.  That’s obviously much more to do with saving costs in difficult economic times than anything related to the lack of available housing.

Nonetheless, I think it might be plausible that high house prices are a factor in extra households not forming — alongside many other factors including more university education, later first child-bearing and others.  But, either way, those households genuinely have not formed and household size is not increasing.  And it also by no means follows that those high house prices are at all related to the fact that those extra households aren’t forming.  The high house prices are a phenomenon of their own.  Once we understand why house prices are high, it can be seen to be very unlikely that any plausible increase in housebuilding could, over any short timescale, reduce house prices.

Even before housebuilding volumes fell after 2007, new housebuilding was less than 1 per cent of the housing stock.  The vast, vast majority of housing sales are of second-hand houses and the determinants of the price of houses are overwhelmingly demand for and supply of second-hand houses.  Some folk have the intuition that “economics is about demand and supply so restricting supply must make a big difference”.  But “supply” in the housing market is, overwhelmingly, not supply of new houses.  It is supply of existing houses for resale.  That sort of supply is largely unaffected (particularly in the short term) by housebuilding.

Instead, housing demand and supply is driven, overwhelmingly, by factors just as wages, job opportunities, employment security, and mortgage credit availability.  And that is true at a regional as well as a national level.  House prices are highest in London and the South-East not because we’ve “run out” in those regions, but because those are the regions with the greatest earning opportunities and thus the highest willingness to pay and lowest willingness to sell to move out.

To see the role of mortgage credit, we can consider the relative movements in the Halifax standardised house price and in the Bank of England’s M4 lending data on outstanding lending secured on dwellings (i.e. mortgages).  By the last three months of 2007, as house prices peaked, the average house was 2.3 times as expensive as in the first three months of 2000.  Over that same period mortgage lending outstanding went up 2.2 times.  As of the first three months of 2015, house prices were 2.3 times their early-2000 level and mortgage lending was likewise 2.3 times its 2000 level.  Mortgage lending and house prices do not move in lock-step — as prices rise rapidly, the volume of outstanding mortgage debt inevitably grows slightly slower (since existing mortgages were secured at lower prices).  (That also illustrates that by comparing mortgage volumes to house prices we aren’t simply measuring the same thing twice.)  But the way that over an extended period the relative movements average out illustrates how important the role of mortgage credit is in driving house price movements.

We talk of how “credit” was easily available in the 2000s and how households “borrowed” from banks and we criticise banks for their “loose lending criteria” in that period and perhaps also criticise governments for “distorting” lending via implicit bailout promises.  But that discussion seems strangely divorced from consideration of what the lending was actually for.  The “loose lending” was, overwhelmingly, “loose lending” to buy houses.  And that “easy money” bid up the prices of houses.  That’s not the only factor in house price rises — employment security, expected income growth and other factors were important as well.  But it was a much more important factor, over that short timescale (in which even if housebuilding had been inadequate to keep pace with population growth, which it definitely was not, as I’ve shown) than changes in new housebuilding volumes, which, being so small could have had almost no impact on prices.

The UK has not run out of houses.  Population growth has not outstripped housebuilding volumes.  And house price rises over the 2000s were almost nothing to do with changes in new housebuilding.  New housebuilding has recently fallen away, and it probably does need to rise in the future.  Furthermore, there are of course many individuals who have personal housing crises for a variety of reasons – mental health, problems getting benefits, poor behaviour with landlords, problems with local authority housing lists, and many other issues.  And there are many splendid housing charities that do excellent work dealing with the personal housing crises that these people face.

But Britain as a whole does not have and has never had a “housing crisis” on any meaningful definition.  People have been talking about this “housing crisis” for more than fifteen years now and yet you may have noticed the conspicuous absence of this “crisis” turning into anything visible.  Have you worked it out yet?  Though some individuals have personal housing crises, the country as a whole (and regions within it) do not have any housing crisis and never did have one.  The idea we had a “housing crisis” was just never true.

Andrew Lilico is Chairman of Europe Economics.