The mysterious tale of two property prices: Nationwide and Halifax tell different stories about house values but both may be right, explains Vivien Goldsmith

WHETHER or not they are about to move house, homeowners are always keen to discover if the value of their property is going up or down.

So it is hardly surprising that the statistics released each month by the country's two largest building societies, Halifax and Nationwide, command close attention.

Yet frequently they tell different stories. And more often than not of late, one bunch has been saying prices have risen while the other has claimed they have fallen.

So who is getting it right - or in some mysterious way, can they both be right?

Over the longer term the figures from Halifax and Nationwide produce a similar pattern, but the month by month reports seem wildly divergent.

One of the key differences between the two sets of figures is that Halifax seasonally adjusts its statistics while Nationwide publishes unadjusted data.

January, for instance, is usually a soggy month for house sales and prices are usually weak. Halifax reflects that in its published figure in an attempt to give a truer picture of what is happening to the underlying market.

So the raw figure for what happened to prices in January as calculated by Halifax was a fall in prices, compared with December, of 1.3 per cent. After adjusting for the seasonal factor, the figure came in as a much shallower downturn of 0.4 per cent.

December and January are the main weak months when the effect of making a seasonal adjustment is that any rise will be plumped up, and any fall diminished.

In contrast, July and August are traditionally strong months. So the effect of a seasonal adjustment then is that rises will be damped down or falls exaggerated. The other months do not vary too much from the norm.

There are other reasons for the different figures produced by the two societies, however. Another key factor is the sample of properties used to produce the figures.

Halifax adds around 150,000 to 200,000 properties to its mortgage book each year, so its figures are based on around 12,000 to 20,000 transactions each month.

Gary Marsh, a Halifax housing economist, says this is large enough to give a true reading.

But because the properties mortgaged by Halifax are not necessarily the same type from month to month, the society adjusts its monthly figures to reflect a 'base' mix of properties fixed 10 years ago. The country's housing stock changes very slowly - it was just over 15 million at that point, with only around 200,000 new homes added each year.

'We will rebase eventually, but there's no real need to do so yet,' Mr Marsh said.

But Nationwide recalculates a different base mix of properties each quarter. Its layman's guide to house price statistics says: 'Simple averages are inherently misleading. If you don't make allowances for the influence of physical and locational differences on house prices, you could well end up doing the housing equivalent of comparing a Rolls-Royce with a popular hatchback.'

Nationwide followed Halifax with a new statistical system in 1989. It refined Halifax's figures by adding more details about the location of the property. For instance, homes may be on a less well-off council estate, in a high status non-family area or in an agricultural area.

It also added in the size of the property in square feet. 'Three rooms in Buckingham Palace is not the same as three rooms in my house,' said Barry Bissett, Nationwide's house price manager.

'We would be worried if, having a different sample, we came out with the same answer. It would mean that adding the extra variables had had no effect. If the differences had been large then, of course, we would not be too happy,' said Mr Bissett.

But Nationwide's sample is smaller than Halifax's - it has just under 1 million borrowers to Halifax's 1.8 million. 'We are more than happy that we have an adequate sample,' said Mr Bissett.

The way that the two societies account for the changing regional mix in properties sold each month also differs. Halifax uses its own figures and believes that its own lending book is large enough to give a true picture.

Nationwide uses national figures from the Department of the Environment to give a regional picture as it recognises that its own lending is biased towards the South and South-east.

But Halifax claims that this methodology can give rise to distortions because Nationwide may have a small sample in one region - say Yorkshire and Humberside - which has a large weighting in the national picture, so a small unrepresentative sample can be given undue weight and skew the results.

(Photograph and graphic omitted)