RPI/CPI User Group

last person joined: 3 months ago 

To foster co-operation and the exchange of information between the ONS, its advisory bodies and other users. Please see 'Aims of the RPI/CPI User Group' in the library for further details.

Report on incorporating OOHPI into the HICP

  • 1.  Report on incorporating OOHPI into the HICP

    Posted 06-12-2018 02:00

    The Eurostat report on incorporating an OOHPI component in the HICP was released on Monday, December 3:

    Due by December 31, it was received by the Secretary-General of the Council of the European Union on November 29.

    The decision made is, of course, most unsatisfactory. I would hope that John Astin, who was head of price statistics for Eurostat for many years, would give us his own evaluation of the report. I may provide my own comments later.

    Prior to the Brexit referendum and even after it, there was always the possibility that the Bank of England's target inflation indicator would be changed in a fundamental way through the back door. Eurostat would rule that OOHPI would have to be part of the HICP by 202? Since the UK CPI is just the UK HICP, housing prices would re-enter the Bank of England's target inflation indicator not because of a change in the remit from the Chancellor of the Exchequer. Now, while that might still happen, the probability it will has shrunken almost to the infinitesimal. Arguably, that's not such a bad thing. That wouldn't have been a very democratic way to make such a fundamental reform. In any case, the UK can't count on Eurostat to save it from future housing bubbles. Britain must save herself by her exertions, and will, as I trust, save Europe by her example.

  • 2.  RE: Report on incorporating OOHPI into the HICP

    Posted 17-12-2018 16:46
    Dear Andrew

    I have never liked the Net acquisitions approach to OOH costs because it does not properly represent the user cost of housing, and am pleased that Eurostat has at least shelved it.

    In my view the only satisfactory approaches to OOH costs are Rental Equivalence and User Cost.   My preference is for User Cost.  ONS has tried using Rental Equivalence but does not have adequate data sources to do so properly, giving rise to a serious underestimation of OOH costs.  There is also a time lag before rents respond which for some purposes is very undesirable.

    As for the BoE inflation target, this does not need to be the same as the overall measure of costs used to estimate inflation.  For the latter I would like to see the User Cost of OOH included.

    For inflation targeting I support a multidimensional approach which would include a measure of inflation excluding OOH costs and also a House Price index.  House prices in the UK are usually the early indicator of coming inflation.

    Happy Christmas.


  • 3.  RE: Report on incorporating OOHPI into the HICP

    Posted 18-12-2018 21:01
    Edited by Andrew Baldwin 18-12-2018 21:05

    Dear Gareth:

    Thank you for your reply. When I posted on December 5 I anticipated a deluge of responses. So far you are the only person to respond. However, I suppose these days people are busy cutting down trees and putting up reindeer and so forth; there may be more of a response in the New Year.

    I like your use of the word "shelved" rather than cancelled for the OOHPI project. Eurostat has been compelling EEA countries to calculate OOHPI countries for some time now. However, there was no announcement that the OOHPI project was being dismantled. In fact, the report concludes that the Commission should "pursue the methodological work" required for inclusion. Presumably, this means that since the ECB doesn't like it that the "OOH index excludes the price of land" they will move to a "net weight, net price" variant after forcing all EEA countries to use a "net weight, gross price" variant in calculating quarterly OOHPIs these many years. If not, it is hard to know what it could mean.

    The HICP is a macroeconomic index, designed primarily to be the target inflation indicator of the ECB, or of national central banks outside of the euro area, although oddly, only the Bank of England, out of all the EEA central banks outside of the euro area, has chosen to use it for that purpose. Eurostat imposed two important rules on an HICP: that interest payments or credit payments be out of scope and that a price index should be based on actual prices, not fictive prices. The user cost approach you favour violates both of these rules; rental equivalence violates the second one. So really an OOHPI for inclusion in the HICP can only be based on the net acquisitions approach.

    For that matter, a strict interpretation of the second rule, which I would certainly favour, means that only the "gross weight, gross price" variant should be used in calculating an OOHPI, since using net prices, i.e. dwelling prices, one is necessarily creating fictive prices. On the other hand, it would even more strongly offend those who don't like the scope of the HICP straddling the production boundary, but so what? The objective of the HICP was never to create a cost-of-living index. If it were, much different decisions would have been made in its construction. And besides being more compatible with the second rule it would also be more compatible with the mission of a central bank in leaning against unwanted asset price bubbles. In its OOHPI manual, Eurostat speculated that besides residential land, asset prices of "works of art, antiques or precious materials" might be treated as in scope. It is a shame that the EC has now chosen to undermine that way of thinking.

    Shaun Richards has already blogged on the EC report at notayesmanseconomics.wordpress.com. He is skeptical about whether a "net weight, net price" variant really can strip out land price movements from house price movements, and with good reason. The Canadian CPI relies on both dwelling and house price indices, and the dwelling price indices are still suspected to be contaminated by land price movements even after many decades of calculating the new housing price index. Nevertheless, Eurostat could always calculate a set of dwelling indices and declare victory, no matter what they looked like. Good for them, if they do, but an HICP incorporating an OOHPI based on a "net weight, net price" variant would be a pretty modest step forward after such a long wait. It is really time for the ONS to go its own way on this issue.

    While I don't favour an opportunity cost approach to OOH in a macroeconomic price index as you do, I would prefer it to the rental equivalence approach, if it were a binary choice between one and the other. Merry Christmas yourself, Gareth, and all the best to you and your family in 2019!

  • 4.  RE: Report on incorporating OOHPI into the HICP

    Posted 19-12-2018 13:02
    Dear Andrew

    Though we disagree on price indices, I judge from your first paragraph that you may share may liking of Joni Mitchell, one of Canada's best exports - and without the aid of a free trade deal.


  • 5.  RE: Report on incorporating OOHPI into the HICP

    Posted 19-12-2018 14:50

    Dear Gareth:

    Glad to see you recognized the reference to Joni Mitchell's "River", Gareth, and that you like her music too. She was born in Alberta (in November 1943, so she turned 75 this year), but grew up in Saskatoon, Saskatchewan, where my parents met each other. "River" was written when she was living in LA, disappointed that Christmas-time there didn't look much like it did on the Canadian Prairies: "But it don't snow here / It stays pretty green." It is a pity that countries that have real winters don't meet together to discuss price index methodology from time to time by themselves. There would probably be a lot more price indices with seasonal baskets if they did, and natural Christmas trees would be a priced item in many of them.

    Best regards,


  • 6.  RE: Report on incorporating OOHPI into the HICP

    Posted 19-12-2018 16:01
    For those who don't know, "River" is one of the songs on the album "Blue" which is widely regarded as one of the best of all time.