Housing affordability index – finally updated (last done in 2004)
Posted by markharrison on December 10, 2006
For those of you unfamiliar with the index, this is a composite that tracks average UK house prices, average UK incomes, and interest rates. Basically, it measures the ability of the average person to pay the INTEREST the mortgage on an average house.
If you just want the headline summary – things aren’t too bad, house prices don’t look overpriced compared to historical norms for this index. In fact, the index is about the level it was in 1983.
If you want to have a look at the index, look here.
If you are interested in the technical nitty-gritty, read on.
The house price data comes from the “Communities and Local Government” website (formerly known as the Office of Deputy Prime Minister.) It is “mix-adjusted”, which means that if one month a lot of big houses sell, and the next month a lot of small houses sell, the statistics take this into account and it doesn’t look like prices have gone because smaller houses sold better than bigger ones this month. (This is, as I understand it, particularly an issue in London where a lot of expensive house sales follow the city bonus cycle.)
The average incomes are from the Office of National Statistics. They are “seasonally adjusted”, but as good a source as any (and better than most.) The slightly annoying thing is that the ONS have changed their website around a lot since 2004, and thus I couldn’t find the older figures. Hence I used the ones from 2004 – which introduced a small error, since I only had the figures on a PDF, and has lost the spreadsheet, it means that figures prior to 1990 are only correct to two decimal places.)
The interest rates are from the Bank of England website. I’ve picked the “quarter close” figures which, again, is a methodology I picked before I ran the graphs. In some years, there was no change, in others there were many wild changes throughout the year, and there is an argument that I should have picked an “average for quarter”. I don’t believe that the different would have been material though.
Right – from here on in it gets REALLY technical. But I’ve always argued that investors should understand economics, so here we go…
There have been some criticisms of this index over the last couple of years, not without merit. The biggest is that by measuring interest payments only, I’m not taking into account the total cost of repaying the mortgage. Obviously, when interest rates are at 14%, the proportion of the cost attributable to repayment rather than interest is low, compared when interest rates are at 3.5%
Hence, it has been argued that what I ought to do is come up with a “total cost of ownership” index, based on amortising the purchase price over 25 years and adding that to the interest. (This would be easy to do – it would basically mean adding 4 basis points to the interest rates). I have run that analysis, by the way, and the figures it comes up with look worse – suggesting prices are where they were in early-mid 1988.
The reason I am sticking by the “interest-only” index is a value-judgement. My thesis is that house prices are set at the margins, not at the averages – and thus properties will sell to the highest bidder looking to buy, not the average man in the street who happens not to be moving. Anecdotally (and this is where I wish I had figures), I am seeing many, many transactions being funded on an interest-only basis, not just from BTLers, but also from Owner-Occupiers. Unless there’s a structural change in the way that houses are typically bought, I don’t see this changing.
The problem, of course, is that this wasn’t the case in the mid-1980s, so we can’t use the either set of figures to really compare. What I believe I need is a set of statistics showing the proportions of transactions done interest-only, interest+repayment vehicle, (and cash-purchase), and to come up with some composite weighting that took into account the raw-interest figures and the interest+repayment figures in an appropriate weighting each quarter. Sadly, I don’t believe that this data-set exists (the growth of interest-only + endowment mortgages means I can’t use the repayment/interest-only statistics.)