In the blog he correctly claims that my case rest on two pieces of evidence:
- the stock of dwellings has grown consistently faster than the number of households since (at least) the house price nadir of 1996; and
- market rents - the price of ‘housing services’ - have grown substantially slower than median household income since 1996.
In the subsequent analysis Professor Cheshire makes four claims by way of rejoinder. One is incorrect. The others are in line with what I’ve been saying. I’m left wondering whether he really disagrees at all.
Claim 1: ‘New house completions, rather than ‘net additions’, is the right measure of new supply’.
TL;DR — no, it’s not.
For reasons that the blog leaves unclear, Cheshire argues that net additions is the wrong measure of new housing supply because it includes conversions and subtracts demolitions. This is an odd argument. You can’t live in a demolished house but you can live in one that used to be an office.* What’s more, I’m not aware of any housing price modelling papers that seek to exclude conversions or ignore demolitions — for the simple reason that it doesn’t make much sense.
There really should not be any uncertainty about this. Net additions is the appropriate measure of new housing supply, as MHCLG makes clear when it says: “‘Net additional dwellings’ is the primary and most comprehensive measure of housing supply.”
Claim 2: ‘Household formation is constrained by the number of houses’
TL;DR — of course it can be, but it hasn’t been
This is a common argument I encounter, but it doesn’t get us very far. The relationship between housing costs and household formation is something that I’ve been talking about since the start of this housing blog series, and gets careful coverage in the CaCHE paper. Naturally, household formation can be constrained by inadequate supply. But since the surplus dwelling stock hasroughly doubled over the past 25 years this constraint has eased, if anything.
Of course, what can really impede household formation is the price of housing services, as captured by market rents. No joy here for supply advocates either though: as Cheshire all but admits, average affordability has been improving. On both key metrics then, housing supply has been sufficient to reduce constraints on household formation. All else equal.
The real constraint has been the paring back of redistributive policies that enable lower income families and young people to afford a home: primarily social housing and housing benefit. The centrality of these policies to household formation is almost always ignored by supply advocates. But housing supply can’t do much to overcome the effects on affordability of, for example, extending the requirement for shared accommodation to single people up to the age of 35.
Claim 3: ‘As people get richer they demand more housing’
TL;DR — correct. The price of housing services tells us if demand has outstripped supply.
Cheshire’s third point is that people want to consume more housing as they get richer. This is true and fairly obvious. It’s common to find that rents, and hence house prices, increase by around 1% when household incomes increase by 1%, all else equal. And as he points out later in the piece, market rents are the true price of housing services, while house prices are merely the present discounted value of the future flow of those rents.
The logical implication of this is that we should expect to see the affordability of housing services deteriorating if supply had been failing to keep up with demand from increasingly affluent households. In practice however, market rents have been drifting down relative to median incomes for the past 25 years. By Cheshire’s own logic there is no evidence that supply has failed to keep pace with rising demand.
Claim 4: ‘Falling housing yields should come as no surprise’
TL;DR — correct. Falling interest rates are the overwhelming driver of rising house price-to-income multiples.
Finally, Cheshire gets to the heart of the issue: the role of falling interest rates. He accepts that housing yields have fallen in line with those on other assets, while rents have been flat. In other words, he is saying that ‘it’s no surprise that house prices have risen due to falling interest rates’. This is certainly progress on his previous rebuttal, where unprecedentedly low interest rates didn’t get a mention.
I couldn’t agree more — in fact that’s exactly the point I’ve been making. Indeed since my paper, two recent pieces from Bank of England researchers have confirmed my finding that the fall in interest rates fully explains what has happened to house prices over the past 25 years. The only odd thing is why Cheshire frames this as an objection.
So Cheshire agrees that falling interest rates have been (at least) a major cause of rising prices. He also essentially now accepts that the cost of housing services has become more affordable in recent decades — but doesn’t explain how his shortage story fits that fact. And to the extent he disagrees on the stock figures, his case is based on what MHCLG says are inappropriate numbers.
In practice, it’s becoming increasingly accepted, as the Bank of England analyses underline, that ‘inadequate supply’ was not the cause of the house price boom of the past 25 years.
There are two more reasonable debates to be had. One is whether increased supply could nevertheless be a solution to the price boom caused by falling rates (although the evidence here, including the government’s own, also seems pretty clear that it could only have a marginal effect). The other is whether it would be a socially desirable use of scarce investment to create the millions of empty houses that would be required even to make a dent in the effects of global economic conditions on house prices.
— — — — — — — — —
*There have been serious quality problems with some of the properties that have been created under ‘change of use’ rules, but even without any change of use properties, the other 212,000 net additions far outpaced household formation rates last year.