From Bloomberg yesterday came the below report stating that house prices in London surged in September after falling the previous month:
London home sellers raised asking prices by the most in seven months in September as a lack of properties for sale and investors looking for safer assets amid financial-market turmoil bolstered values, Rightmove Plc said.
Asking prices rose 2.4 percent from August, when they fell 3.4 percent, the property website said in an e-mailed report today…
Demand in London has been boosted by cash-rich buyers investing in property amid European financial volatility over attempts to avert a Greek default, Rightmove said.
“London’s buoyant property market looks set for a brisk autumn as buyers chase a more limited choice of fresh properties,” Miles Shipside, commercial director of Rightmove, said in the report. “With the continuing turmoil in the financial markets, and the threat of a Greek default, we are seeing a flight to safe assets.”
National asking prices rose 1.5 percent in September from a year earlier to an average 233,139 pounds ($368,700), Rightmove said. In London, Britain’s most expensive property market, prices were up 7.2 percent on the year to 427,889 pounds.
Unlike Australia, which has only one free provider of house price time-series data (the Australian Bureau of Statistics), the UK is blessed with a multitude of providers where house price data can be downloaded for free by the public.
But as with all things in life, greater choice also brings with it greater complexity. And there is a significant degree of variation between the way in which each index is calculated, making comparisons of UK housing valuations across time problematic.
Thankfully, a nice summary of methodology supporting each index is provided here, with each providing different results (see below chart).
For example, the Rightmove index discussed at the start shows mean (average) asking prices when homes are first listed for sale. This arguably makes this index unreliable for measuring price changes over time, since selling prices can change significantly from when homes are first listed. The measurement of average prices instead of medians is also problematic, since they are more susceptible to compositional bias.
In previous articles on the UK housing market and for cross-country comparisons, I have used the Halifax house price series, which measures like-for-like median prices via hedonic regression (in a similar way to RP Data-Rismark). The decision to use Halifax was initially based on the fact that RP Data had used this index in previous market updates (for example, here), and since they are the leading Australian house price data provider, it followed that I should use the same index.
The Halifax time-series shows that UK and London house prices have fallen by around 20% nominally since prices peaked in September 2007:
Nationwide also provides a house price index that measures median like-for-like house prices via hedonic regression. But this index paints a different picture of UK house prices. According to this index, UK prices have fallen by 11% in nominal terms since the peak, whereas London prices are down only 2% nominally:
Finally, there’s the Financial Times house price index, which has the benefit of using the broadest sample of home sales (from the the Land Registry) but suffers from measuring means (averages) instead of medians and does not adjust like-for-like sales (via hedonic regression). This index shows that UK house prices are nominally 6% below their peak, whereas london’s prices are flat:
Putting these indexes together paints a somewhat confusing picture of UK house price movements:
I guess the lesson here is to tread carefully when interpreting any house price data, something lacking in the Bloomberg headline.
