The 3.5 Times Earnings Debate

[spline]

The central idea is that ODPM figures show that, apart from when well into a bubble, both OOs and FTBs borrow at about 18-20% of gross earnings, irrespective of IRs, and cycles (excepting the bubble peak when the fear/greed thing kicks in). Now it’s easy to take historic earnings and IRs and convert them to the value of a 25-year repayment loan. Then use ODPM data on price:earning and loan:earning to gauge the ratio between the average house price and the average loan - this gives a price/loan of around 1.5 to 1.6, and hence the p*. If prices go above this then people are stretching the repayment:earning ratio – this, again from the ODPM data looks like a Gaussian hill, so just hack one in with appropriate scalings. However there does seem to be a time lag between the price response and changes to the p*, so I’ve done this as a simple linear damper with a characteristic time of about 1.0-1.5 years.

The YoY comes out like this – post 2005 predictions depend on how much of the present peak is split between the ‘bubble’ and IR ‘lift’ components. Although I suspect that with a bit more care in the tuning the predicted YoY in 2007 might be slightly more negative.

Edit: the 1980 miras handout and, crucually, its removal added to the unsustainable excess price which then corrected down to a p/e of 3.2, i.e. lower than if it hadn't been removed.

All I've done here is to start with a simple core idea and explore where it takes us.

Edited September 18, 2005 by spline

[frugalista]

The UK is not the Soviet Union. If lenders are willing to lend at 6 or 7 times the income let them do so. This is a free country.

<{POST_SNAPBACK}>

The UK is not the Soviet Union but it is not the Wild West either.

Caps on lending multiples would not equate the UK to a command economy. The UK can be a mixed economy, like the rest of the EU. Let the market run most of the economy, but regulate those bits which benefit society by being regulated. Massive swings in house prices does more damage than good surely.

frugalista

[spline]

The technical details:

- Idea of (almost) invariant repayments:earnings M/E gives Ps(t) as fn(E(t), R(t))

- actual prices bubble above this, so add B(t) excess (use ODPM data as guide)

Notation:

P average house price estimate (blue, purple lines)

Ps 'steady state' house price estimate (P* on graphs, no bubble excess, green line )

M average loan repayment (per annum)

E average earnings (gross, per annum)

L average loan advance

R mortgage Rate (BoE base + Roffset), with Roffset = +1%

B bubble (or excess) factor, (B=0, no bubble)

T time (years)

Equations:

P/E = (1 + B )*Ps/E

Ps/E = A*L/E, with A = 1.5

L/E = annuity(M/E, R, 25 year)

M/E = 0.19

R(t) = historic IR rates (time series) BoE + offset

E(t) = historic earnings (time series) ONS

P/Ps = overprice ratio = (1 + B )

B/E = (P–Ps)/E -1

Two bubbles – B = B1 + B2, Gaussian hills (at 1989 and 2004)

B1 = 1.75, B2 = 0.58

Bi(t) = Bi*exp((t-t_i)^2/Tb^2) Gaussian bubble, tb=peak year, t=years

Tb = (sqrt(3), t<ti; 2, t>= ti) Bubble time scale, slight up/down asymmetry.

t1 = 1989, t2 = 2004 (but with a slight back-shift when using the damped response)

Market response (linear damping) – not essential, but tracks the trough oscillations quite nicely when using it, otherwise simplify using P = Ps*(1 + B )

Ta*dP/dt + P = Ps*(1 + B )

Ta = 1.5 year

Pre-1990 tax advantages ignored (these boost P prior to 1990)

the removal of this boost pre-loads the 1989 bubble ...

Edited September 18, 2005 by spline

[eric pebble]

I think this one of those things where it's just too late now.

If only they'd enforced this restriction a few years ago (and restricted btl lending) we'd probably be fine now.

Also when the mortgage markets were liberalised to take female earnings into account if only they'd kept the joint earnings limit low!

Too late now, a clampdown would be electoral suicide at this point.

<{POST_SNAPBACK}>

Yes - the Moneylenders have had their way -- and they will get more and more avaricious -- and more and more powerful...... Really really sad - and bad.

[Karen]

This has been raised many times on HPC, that historically the earnings ratio in the UK is around 3.5 times salary. Higher in London, lower up North.

So it has been suggested that if the government could restrict borrowing to this level (ie proof of income from Inland Revenue). This would stop Joe Public getting himself into the brown stuff.

However it's not very fair in a low interest rate environment:

Joe Soap earns £40k, and rents a London flat worth £240k.Six times gross salary.

Rent costs £1200pm, £14400 a year.

So the place is generating a 6% gross rental yield.

Now Joe can get a mortgage at 4.75%, so it's going to cost him 11400k a year to buy (IO mortgage).

Now currently Joe can go to a mortgage company and borrow 6 times earnings.

By doing this he can save himself £3000 a year.

He has to do the maintenance, but he's quite handy and doesn't mind the odd bit of painting.

Now what do you guys think is fair?

1) Allow Joe to buy the place, save himself some cash each month, and he's happy knowing it's "his" place. It may fall in value, but that's a risk he takes.

2) Enforce the 3.5 times rule, Joe can't buy, and is forced to rent. Professional investor's buy, and make money off Joe.

If you were the goverment, what would you do?

<{POST_SNAPBACK}>

Honestly I like the 28% rule better (well, honestly, I think 20-25 is even better, the 28% rule is where people tend to get in trouble, and the bank has to forclose. The bank does not care how often you eat rice and beans). Meaning your payment should not be more than 28% of your monthly income. I also favor 30 year fixed loans. Here in the US IO loans generally only go for 1-5 years, So after that time your payment goes up as much as 80%.

True to life example in my area. Houses worth 350K rent for about 1000K a month.

Now assuming you had the 70K (and I would put an average income in the area at 60K, so the 20% down would be more than 1 years income)to put down, your payment on 350K at 5.5% for 30 years would be $1,589. Do a no down deal and your payment is 1,987.

Why would anyone pay 1,987 a month when they can rent the place for 1,000?

This is how I know we have a bubble in the US and not a housing shrotage. If it was really an issue of not enough housing rents would be going up just as fast as housing prices.

And like you said, it may fall in value. What if it falls in value, and Joe needs to sell?

So for example, Joe buys 350K house with an IO loan. He can make the payments. After a year he needs to sell because he lost his job. The house is now worth $325. Where will Joe get the extra 25K?

[camem']

I suspect he means loan-to-value ratio (ie the loan as a percentage of the property's appraised value):

http://www.cml.org.uk/servlet/dycon/zt-cml...l_Table-PR2.xls

See the "FTB's percent advance" and "Movers percent advance" columns.

I find these statistics interesting - In 1980 the median income multiple for FTBs was 1.73. Even with a very high recent 'FTB salary' of £36k (http://www.moneynet.co.uk/press-release/press-release-3feb05.shtml - only because other FTBs are priced out I suspect) this implies an 'average FTB house' for 61k. Is there a reason why we can't see this again (apart from general rioting/revolution/civil war because people's houses have fallen 67% in value) ?

The other one is the 'percent advance' - in 1980 the median (most common) FTB could afford 16% deposit (suggesting some had much more) . From today's prices (average 180k, let's say) that is around 28k. Imagine the *median* 20-30 something owning 28k deposit now (rather than owing 28k on cards)...

good ammunition against those who say the young have never had it so good.