Wednesday 29 December 2010

House prices vs earnings, England and Greater London

Sources:
1. Average price for a semi detached house in each local authority area from the BBC
2. Mean gross annual pay from Table 8.7a from the NSO
3. I then bunged the two data sets into a spreadsheet, deleted local authorities for which I had earnings but not house price (or vice versa) and did a couple of X-Y scatter charts, click to enlarge:
---------------------------------
That doesn't look like much of a correlation, but Excel tells me that the correlation is 0.75, which is high enough for these purposes.

The line of best fit is House price = (earnings minus £6,750) x 10.53, which makes sense. People have to pay for the bare minimum of existence first and the bulk of whatever is left over goes on housing. With an average marginal tax rates of about 50%, and average mortgage interest rates of 5% (or average rental yield of 5%), for a single-earner household, nearly all the extra income you can earn by moving to a higher wage region (assuming you want the same standard of housing) goes in extra rent or mortgage payments.*

The bulk of the fluctuations can be explained by how desirable an area is to second-home owners, for example an average semi in Torridge in Devon costs £195,583 and average earnings are £16,980; and an average semi in Colchester costs £203,921 and average earnings are £30,466.
-------------------------------
The figures for Greater London are, as ever, completely off the scale, but the correlation between house prices and wages is even higher than for the rest of England at 0.86.

The line of best fit is fairly nonsensical, it approximates to House price = (earnings minus £24,195) x 45. This would suggest that Londoners need two or three times as much money (after tax) to cover what they consider to be essentials as do the rest of England, but once those essentials are covered, they are prepared to spend more than the rest of their income on housing - unless, of course, rental yields or mortgage interest rates are very, very low (about 2%).

In a house price bubble (which is popping slowest in London), people see hoped-for capital gains as a source of cash income and deduct them from interest costs (or add them to rental income), i.e. it is the same whether people expect house prices to rise 3% and interest rates are 5% or whether people expect house prices to stay flat and interest rates are 2%.

Other people for whom the effective interest rate is 2% or less are people who have paid off their mortgages - they could in theory sell up and start renting, so the cost of being an owner-occupier is the interest foregone on the theoretical pile of cash. As it is difficult to earn more than 2% after tax on a savings account with a UK bank, their effective/notional interest cost is thus 2% or less.
-------------------------------
Nonetheless, the overall picture is, as we would expect, that there is a fairly high correlation between earnings and house prices (taking England as a whole, including Greater London, the correlation is 0.79), so replacing all taxes on incomes and output with a flat tax on land/building values doesn't seem particularly radical to me, i.e. at my suggested rate of 'about' 8% of current averaged out selling prices in each area, and assuming 1.5 earners per working age household, eighty or ninety per cent of people would be paying the same or less in publicly collected tax (i.e. those households with a house price-to-gross income ratio of seven or lower).
-------------------------------
* This is as suggested by Ricardo's Law of Rent, because labour within the UK is mobile but houses aren't. So if people expect the same living standard wherever they are and net wages in Town B are £10,000 more than in Town A, people will move from Town A to Town B, thus increasing demand for homes in Town B, thus hiking rents and prices in Town B. This process continues until rents and mortgage payments in Town B are £10,000 more than in Town A, at which stage there is no advantage in moving from Town A to Town B and the whole thing reaches a new equilibrium.

15 comments:

Bayard said...

Which county has the worst house price to earnings ratio, i.e. the least affordable housing? I bet it's Cornwall.

Mark Wadsworth said...

B, the most extreme is Torridge in Devon (as mentioned). Cornwall has average earnings £20,000 and average house price £198,000.

As it happens the closest to the average for England excl. Greater London is Fareham in Hampshire, with earnings £26,000, average semi £206,000.

Lola said...

Yes, yes, yes - I know all this. Now, as previously asked, where do I sign up for the fix?

Mark Wadsworth said...

L, where to sign up is the big problem. Which is why I tend to go on a bit about this, just to show the huddled masses that it is a perfectly sensible and realistic way of running things.

BlackRaven said...

1) "As it is difficult to earn more than 2% after tax on a savings account with a UK bank, their effective/notional interest cost is thus 2% or less."
It is a common mistake for UK households, however it doesn't make sense to compare the interest earnt on a savings account to opportunity cost for owning a house. There is an enormous difference in liquidity of the assets; I don't know but the average time to sell a house is going to be more than a few months I imagine, plus you have large transaction costs.

2) As you have stated before, the effect of such a measure as LVT would be to equalize property prices across the country. That pressure to disperse dense areas of commercial activity from a LVT is not costless or calculated in your modelling and an important unintended consequence.

Mark Wadsworth said...

BR:

As to 1), agreed. But people do factor in the cost and hassle of moving when making the decision whether to downsize or sell-to-rent.

As to 2), LVT would tend to equalize residential land values; but in the UK we have Business Rates on commercial premises (i.e. "dense areas of economic activity") which are pretty close to LVT and which would remain the same for the time being (or else the maths gets too complicated).

So seeing as "dense areas of economic activity" would benefit enormously from income tax, VAT, corporation tax etc being scrapped, there is no reason to assume that anything terrible would happen, is there? Derelict sites will be brought back into use (hooray!) and that's about it.

Alice Cook said...

very nice!

Steven_L said...

Not much correlation, but then there's loads of mugs like me who want to live in cool places like Brighton and get paid a pittance.

Mark Wadsworth said...

AC, ta.

SL, if employers in Brighton know that young people consider it a 'cool' place to live, then they know that they can offer slightly lower wages than otherwise. Add on the increase in prices caused by second homer effect and there's yer discrepancy.

Lola said...

Brighton? Isn't that where the fragrant Caroline Lucas is an MP? She's a Greeny I think, but luckily she's reasonably attractive....

Old BE said...

"The line of best fit is House price = (earnings minus £6,750) x 10.53, which makes sense."

Which means that I must be an outlier, because my home cost half that predicted by that formula and I live right in the centre of London and don't earn a spectacular salary.

Robin Smith said...

This does not match observed fact. People pay rent/interest first, then buy essentials for living.

If they cannot afford rent they move or take benefits.

Earnings are what is left after rent is taken out. The law of rent.

Please keep it simple. People are stupid and selfish right now. We don't need more work.

Mark Wadsworth said...

L, highlight of her career was when David Campbell Bannerman offered to stab her in the face.

BE, that's the ratio for England and relates to semi-detached houses. Try your calculation again assuming you owned a semi, and try the formula for London (wages minus £19k) x 45.

RS, well, exactly not.

In economic terms, people pay essentials first, and the balance goes to rent. If earnings are only enough to cover essentials, then rents are +/- nothing. That is the theory and that is what we observe in practice.

As a matter of law, if you are strapped for cash and have the choice between going hungry for a week or being evicted, you'd prefer the latter in the short term; in the medium term you would find a cheaper place to rent.

neil craig said...

Not much correlation then between the actual cost of building the things which should be similar everywhere (ok some increase where wages are higher or transport infrastructure makes moving materials more expensive).

This is what we would expect if housing costs are almost entirely caused by goverbnment monopoly power. The exception where there is extra demand because of people after holiday homes is the exception that proves the rule.

On the other hand we should expect that in any council area whose "planning" dept believes in the free market more houses would be allowed to be built & costs reduced. I assume this means that councils are, in reality, entirely prevented from doing so by Westminster, Brussels etc.

Mark Wadsworth said...

NC, agreed - except I doubt whether 'the government' (in the narrower sense) gives two hoots about how many houses or roads are built, it is politicians pandering to NIMBYs and Greenies who prevent houses and roads being built.

In any event, we've got plenty of houses, the problem is they are allocated totally inefficiently, which LVT would sort out most admirably.