A case of lies, damned lies and house price indices?

How accurate are house price indices? They are meant to illustrate market fluctuations and give a guide to prices yet they vary according to source. The Halifax and Nationwide Building Societies have both recently been accused of misleading by exaggerating price rises in order to build up the market. But is this fair?

Homeowners in the South-east reading the latest Halifax report will have been delighted, if perhaps disbelieving, to find that their properties have risen in value by as much as 20 per cent over the past year. Those in the East Midlands may not be so happy with a 3 per cent rise.

But where do these figures come from?

The Halifax statistics are based on its own mortgage offers, statistically adjusted on a scale stretching back to 1983 when the index first started in order to remove any anomalies. This process is called mix adjusting. The Nationwide index is also mix adjusted, looking at the pattern of lending over the past four years.

The Land Registry started its own price report in 1995 and uses all sales of freehold and long leasehold (over 21 years) property, whether or not the purchase was supported by a mortgage. Sales which do not represent market value, for example ex-council houses sold at discount, are excluded.

But how accurate are they? At the Bradford and Bingley, the last few months have seen an increase in the number of new mortgages and this has tied in with price indices. But this is not always the case.

"Price indices obviously have their uses," says Chris Holland of Bradford & Bingley, "but it is important to draw the line as to how far one wants to rely on them."

An index showing a dramatic rise may, for example, bring hope to a vendor who will then refuse to budge on a lower sale price. But just because indices indicate strong price rises, they give no explanation of why this is happening. In the South-east for example, price rises are being driven by a shortage of property. This indicates that people are buying and selling only where strictly necessary. As soon as more people feel confident in the market again, distortions in house prices will ease off. "We expect substantially more existing homeowners to return to the market next year," says a spokesperson for the Nationwide, "resulting in more properties for sale and unlocking the severe log-jam seen over recent months."

Those who compile indices defend themselves against charges of inflating prices. Press reports and estate agent hype are more likely to influence prospective buyers and sellers than price indices, they say. And unlike estate agents, as Mark Hemingway of the Halifax points out, "it is not in our interests to talk up the figures. We are simply reporting what is actually happening in the market."

Alan Oliver of the Nationwide says: "If Nationwide and Halifax indices are placed side by side, the trends over a long period of time will closely mirror, and certainly complement each other."

There is less discrepancy between sources than one might suppose. In the last quarter, a semi-detached house in the West Midlands was worth pounds 53,397 according to the Nationwide, pounds 59,677 by Halifax calculations, and pounds 54,688 according to the Land Registry. Both the Nationwide and Halifax reported an average increase of around 8 per cent during 1996: the Land Registry 7.45 per cent. Strong price increases of 10 per cent or more are widespread only in parts of London and the Home Counties and these areas are mainly responsible for the sizeable national increase.

The Halifax covers one-fifth of the mortgage market, providing a large sample for its figures. As it is used for government statistics it has been investigated thoroughly and is generally regarded as an accurate reflection of the market.

Both the Halifax and the Nationwide report at the end of each month and are better at indicating the current picture in the housing market than the Land Registry, which will be up to three months out of date due to the time registrations can take. There are other drawbacks to the Land Registry indices. No adjustment is applied to reflect seasonal or other factors. For example in one month most property sold may be two-bed semis in cheap areas, the next month the market may be dominated by 4 bed-semis sold in more expensive areas. An accurate index needs to build in the ability to cope with distortions in the market, and to iron out inconsistencies.

But as Alan Pemberton of the Land Registry points out, "Ours is not an index and our statistics cannot be compared to the Halifax and Nationwide indices. We provide our information as a public service and, we hope, as a complement to theirs."