Blogs | September 11, 2009

Have house prices fallen more than you thought?

The European Central Bank (ECB) estimates of house prices look simple enough. The ECB says, for example, that house prices in the euro area rose by 0.8 per cent last year – the most recent period for which they have figures. But if you don’t believe this, you’re right. This is because house price indices can be very misleading in a slump.

How can the figures be misleading?
Most indices measure the prices at which real estate changes hands. But in a slump the worst homes tend not to sell at all, so they don't show up in the indices. This highlights a big difference between house price indices and share price indices. The worst performing shares are included in stock market indices, unless the firm goes bust. But many of the worst performing houses do not figure in house price indices.

Comparison of house price falls in the UK (Year to June 2009)

All house price measures are not equal
Some index providers – such as UK mortgage lenders Nationwide and Lloyds Banking, Spain's Tinsa or the German Federal Statistics Office – try to solve this problem by adjusting their indices for variations in quality.

These so-called "hedonic indices" control for factors such as the size or number of bedrooms in a house, or how many parking spaces or bathrooms it has. Such adjustments, though, while an improvement upon the ECB's indices, do not go far enough. Anyone who's tried to buy a house knows that there's huge variation even among, say, three-bedroom detached houses in a particular town. In a slump, the worst such houses won't sell, so even quality-adjusted indices will understate the severity of the downturn. There is, though, a better way to build indices. This is to track the prices of the same houses when they sell and when they re-sell. As long as the few houses that have new extensions are offset by others that have deteriorated, this allows us to adjust perfectly for variations in quality. This "repeat sales" method is used by S&P's Case-Shiller index in the US, and by the Land Registries in the UK and Netherlands.

These indices show bigger drops than those which don't adjust so well for quality. In the US, the Case-Shiller index for 20 main cities fell by 17.1 per cent in the year to May, whilst the National Association of Realtors' index of median prices for single-family homes fell by 15.6 per cent. And in the UK, the Land Registry's index fell by 14 per cent in the year to June compared to falls of 9.3 per cent in Nationwide's index and 12.3 per cent in Lloyds Banking's. There is, however, a problem even with repeat-sales indices, as Yale University's Will Goetzmann has researched. Imagine sellers were to refuse to sell houses at prices below those which they had paid. We would then never see a fall in the repeat-sales index – we'd just have very few transactions. Of course, few sellers have such a rule, and fewer still can stick to it. But the point remains. Sellers' reservation prices – thresholds below which they refuse to sell – can, says Professor Goetzmann, cause observed price indices to become "entirely spurious."

What it all means
The message here is simple. House price indices can under-state the extent to which prices fall in a slump – meaning the measures can cause us to over-estimate the stability of the housing market. Maybe housing is a riskier investment than you think.

EconomicsHousingHouse prices

Chris Dillow
Chris Dillow

Investors Chronicle writer and economist