House price statistics flop

Today’s pre-launch of the new house price index from ONS was a disaster. The proposals lacked ambition and the new statistics (based on Land Registry prices) could well disappear next year when the Land Registry is sold. “We’re not going to speculate” about that, said the ONS, who didn’t know who would own the statistics.

There was an article published this morning on the ONS website setting out the detail of what to expect. This followed a consultation in 2014. The presentation in London (part of the “Economic Forum” series of meetings) added little except a fair dose of disappointment and scepticism.

What’s planned? Both the existing ONS house price series and Land Registry series will cease later this year, to be replaced by this new joint index, the details of which are in the article.

In a number of respects the new series is good news:

  1. Neither existing index is fit for purpose. The ONS series is based on a sample of mortgage data (so excludes cash transactions, and is just a sample). The LR series is based on “matched pairs” or repeat sales so excludes new build, rarely transacted properties and is (probably) overweight properties that have been bought, done up and sold.
  2. One official series is a bit less confusing than having two. It is clearly odd for the government to be publishing two such divergent series.
  3. The potentially very misleading arithmetic mean is being ditched in favour of the geometric mean. (Try this for an explanation if you want one!) A few very high value transactions will be less distorting to the average in future.

But those gains are outweighed by some minor concerns, a massive lack of ambition and one huge risk.

The huge risk is that this joint venture between ONS and LR is likely to hit the rocks as soon as it’s launched when the LR is, as widely expected, sold off next year. (The government has just published its consultation on this.) The new series is based on LR data and the index will be published by them. That is unsustainable when the LR is a private sector body. There seems to be a very real risk that the UK will be without an official index in a year’s time! The ONS “no comment” is not enough.

This “sell-off” factor alone means this plan to replace the existing indices must not go ahead until there is clarity. The timing of the launch was unfortunate but either UKSA/ONS needs to get a guarantee of continued access to the data or stop this project until the fate of LR is decided.

Other issues that raised eyebrows at the presentation included:

  1. The relative lack of detail that will be published in the early releases of data.
  2. The very clear resource constraints that the house prices team is under.
  3. Halfhearted attention given to weighting. The new UK HPI is to be mix-adjusted to allow for the fact that different houses are sold in different periods by annually updating the basket of properties according to what was sold in the previous year. This seemed odd as most people present wanted the weights to be based on the stock of housing not transactions. (You’d never create a stock price index weighted according to how many shares were sold, as one person put it.) The ONS said the reverse was the will of those who responded to the consultation. Could both not be produced?
  4. Why no median data?
  5. The lack of information about the quality of the new data being used for hedonic adjustments, notably the property size data, was striking.
  6. Confusion over the new data access from VOA. Some data seems to have been obtained but not most of what could have been, and its quality is uncertain.

Oh, and someone asked how this data might relate to changes to the CPI. He got the answer: “We’re not taking any questions on CPI.”

A good many attendees were saddened by the lost opportunity implied by these proposals.

The ONS is under pressure to embrace administrative data sources, data sharing, data mashing, adding value and big data – in other words innovating. This pressure was redoubled following the publication earlier this month of the Bean Review. Add to that the nation’s obsession with house prices (and the economic significance of property prices and the active Treasury policy-making) and there was a prime opportunity here to engage in innovative work. New, good quality, genuinely informative data would have fed an appetite among users that would have made the ONS millions (yes, quite possibly) of new friends, reaching an audience way beyond its normal rather narrow group. That’s the opportunity that has been missed.

The ONS could have, as part of its house price suite of data, collected information on and published statistics on the following:

  • Energy performance certificates
  • Council tax band
  • Whatever data the LR has, for example, tax relief claimed, single/multiple buyers, NI numbers, corporate buys etc
  • Any insight gained from links to the micro data from any other surveys such as the English Housing Survey
  • The data on the HMRC Stamp Duty form

The data above is existing data held by ONS or other public sector bodies, but the ONS could also have advocated small changes be made to the data collected by those bodies to make the resulting information more useful. Perhaps in due course links could be made to the address list and rental data? Even some private sector bodies might be tempted to join forces and share data.

None of that would be easy or instant but the ONS could have set out a vision, of which today’s announcement was just a modest first step. By my reckoning, the consultations on this new index started in 2002 (not in 2010 as said in the presentation) so waiting a few years to deliver this would be fine. Such new and in depth statistics would really add value and knowledge. Instead we got a rushed job.


A note on the ONS website.

At the start of the presentation we were told that details of the new data were published this morning on the ONS site. I couldn’t readily find it. It was not on the ONS home page. I clicked on “H” (in the spelling box) and got to the regular house prices release from a week ago (and various other statistics beginning with “H”). I failed to find it with “house prices” in search. Eventually I found it in “news”. I tried to sign up for “news alerts” of news (as opposed to stats releases) but it doesn’t seem to be possible.

And I can’t find “Economic Forum” on the web site – or by using search.

3 thoughts on “House price statistics flop

  1. Re the website you are right – we could have done a better job sign-posting the release. It was indexed by search but it needed some manual intervention to make it more visible with the search rankings and via browse. It is something we will get better at.
    As for finding the Economic Forum on the site – that would be challenging as it isn’t there yet – I met the team last week and we are rebuilding the section including the presentations archive.

    Like

    1. Thanks for the comment. I am not sure what “It was indexed by search but it needed some manual intervention to make it more visible with the search rankings and via browse” means but I’ll take your word for it! ……. as for the old content that is still valid or sought but not loaded up, that seems to be a bigger issue. I do think that more should have been there at the launch of the site and the “Economic Forum” stuff is/was important. Anything about Bean, methodology, innovation etc is highly topical.

      Like

      1. Just means someone should have checked that it was on the top of the first results page and if not tweaked it 🙂

        As for content migration decisions had to be made or we would never have been in a position to launch – some things like Bean related content we should have anticipated (and are now rectifying that) but elsewhere it was always going to be demand led and we made decisions based on the usage analytics / search terms we had – always aware that we’d need to be flexible enough to respond post launch.

        Like

Leave a comment