There are few British obsessions quite like our fixation with house prices, but which of the many indices that measure home values can you trust the most?
This question has been brought into sharp focus in recent months due to mounting criticism of the Halifax House Price index for its wild swings recorded in 2019.
It began the year showing a 2.9 per cent month-on-month dive in prices in January, but then turned full circle, soaring to a 5.9 per cent jump in February.
However, it then slumped back to a 1.6 per cent drop in March – only to return to a 1.1 per cent rise in April and a 0.5 per cent increase in May. Those figures are out of kilter with other, more benign indices.
The Nationwide House Price index has only gently swung this year. Until the end of May, prices had not fallen by more than 0.2 per cent month-on-month, while they had not risen by more than 0.4 per cent.
The government’s UK House Price Index initially showed values fell by 0.8 per cent month-on-month in January and February, followed by a 0.2 per cent drop in March, but it then recorded a 0.7 per cent rise in April.
Other big indices from LSL Acadata and Rightmove have reported much softer trends than Halifax.
After May’s Halifax data was published, Pantheon Macroeconomics chief UK economist Samuel Tombs tweeted: “I want some of whatever the statisticians at Halifax are smoking.”
A few months earlier, after February’s huge rise, he had said: “Very hard to take Halifax’s data seriously. Prices up 5.9 per cent m/m [month-on-month] in Feb – I do not think so. Review your seasonal adjustment methodology!”
EY ITEM Club chief economic advisor Howard Archer calls the Halifax index an “outlier”, while other economist firms have similar concerns.
Others in the mortgage market are unconvinced, too. John Charcol senior technical manager Ray Boulger says: “The Halifax index must have become an increasing embarrassment to Halifax over the past few months.”
Like Tombs, he says a key problem is what he calls the “dubious” seasonal adjustments to iron out the fact prices tend to be higher in spring and summer when more buyers are active.
Boulger adds: “Now one can only access the real figures before any adjustment with a subscription, and there is little point in subscribing to see figures of such limited value.”
Halifax insists its data is robust but adds that it is “reviewing the methodology to see if it can be improved, and we expect to make some changes later this year”.
Understanding the data
To help work out which is the most useful index, it is first important to understand the data behind the big players.
The long-running Nationwide and Halifax indices derive from their mortgage businesses only, with the price at the approval stage used. Therefore, they do not necessarily record the real sold prices.
The Land Registry says Nationwide bases its figures on about 12,000 approvals a month, while Halifax’s index is reportedly based on about 15,000 approvals, though neither lender would reveal the true figure.
Some argue the Nationwide index is skewed towards the south, with Halifax focused more on the north of England, though both lenders deny this.
The UK HPI, which only launched in 2016, is an amalgamation of data from the Land Registry, Land and Property Services Northern Ireland, the Office for National Statistics and Registers of Scotland.
It is based on actual selling prices of about 100,000 properties a month – which it says accounts for every UK transaction – but data for a particular month only tends to get released six to seven weeks after the end of the period in question, so there is lag.
LSL Acadata measures the price of all properties sold in England and Wales using Land Registry data. As well as using a smaller sample size to the UK HPI by virtue of not covering the whole country, it also has a different way of calculating the figures, as it does not use compounded averages like the UK HPI does.
Meanwhile, estate agent aggregator Rightmove collates advertised prices in England, Scotland and Wales from its portal – which it says often numbers more than 100,000 prices per month.
The Land Registry says the Halifax and Nationwide indices suffer comparatively to the other big datasets because they track far fewer properties.
Naturally, they do not record any of the up-to-40-per-cent of transactions where no mortgage is taken out.
In a report on house price indices it published in December, the Land Registry said: “If trends in cash and mortgage sales differ, this may lead to biases in measures that exclude cash sales. Other things being equal, statistics based on a larger,
unbiased number of transactions are likely to produce estimates closer to the true value.”
London & Country director David Hollingworth favours the UK HPI, as does John Charcol’s Ray Boulger. Both highlight the simple fact that it records actual sold prices, unlike some other indices.
“It is arguably the most accurate [index], as it is based on done deals,” Hollingworth insists.
“However, there is therefore lag, as it can take time to complete a purchase.”
Boulger has a similar view, though he also has time for the Nationwide index. He says: “I find a combination of the UK HPI and Nationwide by far the best combination of indices to get a good feel for house price value changes, in addition to first-hand reports from the coalface.
“I am only interested in indices which provide actual prices, as opposed to the manipulated, or seasonally adjusted, ones,” he adds. “When the UK HPI was launched, the government very sensibly – after doing its research – decided the monthly statistics should focus on actual figures, although it does also publish seasonally adjusted figures.
“Although Nationwide’s press release also focuses on seasonally adjusted figures, it helpfully provides a record of actual prices and so it is very easy to work out the real price change.
As Rightmove data is based on asking prices, I consider this has very limited value on actual sale prices.”
Of course, the headline figures from all these data sets have little relevance for a broker or their client when trying to understand the local market in question. While some indices also produce regional figures, even they often fail to have the granularity a buyer requires, given that price fluctuations can occur from one road to the next. It is possible to track local sold prices road-by-road on publicly available tools such as Rightmove and Zoopla, but with lag of a few weeks or months.
“National figures make for good conversation pieces and can be useful in outlining the overall mood, but recent years have shown how much variation there can be in regional markets,” says Hollingworth.
Just how much variation there will be across the board will no doubt be pored over by thousands of economists, brokers, buyers and more over the coming weeks and months, as has happened for decades. How trusted the Halifax data is over
time by those same individuals remains to be seen.