RPI/CPI User Group

last person joined: one month 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.

Topic: National Statistician's statement on the future of consumer price indices

1.  National Statistician's statement on the future of consumer price indices

Posted 10-11-2016 10:43

Dear all,

Just to let you know that the National Statistician published a statement today, covering the future of consumer price indices. This follows on from his letter to the Chair of the UK Statistics Authority in March.

Thanks,

James (ONS)



2.  RE: National Statistician's statement on the future of consumer price indices

Posted 10-11-2016 15:26

Dear all,

You will have seen that earlier James posted a link to John Pullinger’s statement on the future of consumer price inflation statistics. In the statement John says that we will be including Council Tax in CPIH from March 2017. We will also revise the Council Tax component back to 2005 so that the time series is consistent. This gives us an opportunity to revise the OOH weights, which will make CPIH as up to date and consistent as possible when it becomes the preferred measure of inflation next March.

For that reason we have decided to update the weights analysis we have prepared using the most up to date data, and to extend it to also include an analysis of the impact of including Council Tax in CPIH. We are aiming to release this analysis alongside the December consumer price inflation bulletin.

I hope that this clarifies the situation on weights. If you have any questions about this please do get in touch.

Kind regards,

Chris

Prices Development




3.  RE: National Statistician's statement on the future of consumer price indices

Posted 11-11-2016 09:43

By the time of the December release it will be the best part of six months since I made my estimate of the weighting for OOH based on the data released by the national accounts.

Could I ask whether the most recent weighting figure could be confirmed for us now so that if any of us are not assuming a figure close to that of the ONS we can start working out why.

We have been deliberately stalled on this for the best part of 6 months.

------------------------------
AL

Waiting for them to sort out housing in the CPI since 2011

http://www.localisewestmidlands.org.uk/2011/alternative-inflation-report-coping-with-an-inflationary-global-economy/



4.  RE: National Statistician's statement on the future of consumer price indices

Posted 12-11-2016 09:37

Making the CPI-H the central index for inflation any time soon is very big talk from an ONS that seem so strangely unenthusiastic in letting us see their new weights for this CPI-H.


Pullinger's credibility is eroding very fast. These grand statements he puts out never define what is going to happen. It was one of these statements that told us they were going to get the CPI-H ready for re-designation by this Autumn just gone. None of this is of course referred to in his latest one.


I do not know whether he has a clue what is going on or just puts his name to these announcements drafted by the real people who are making a mess of all this.


I do worry that to be saying this about the CPI-H he is prejudicing the issue when he is going to be the person making the recommendation to the UKSA. And he is also a member of it.


I did put some time into trying to get the Select Committees to look into what was happening on the reform of the inflation indices and was getting somewhere on that until Dilnot announced the so-called Johnson Review. I think I will start trying to get them to begin to look at this issue again.


Any CPI-H that is re-designated in the foreseeable future will not be very robust under scrutiny, and will help tattify the Johnson Review which I increasingly think was a waste of time.


These comments are fair comment in the public interest about performance in a public office, but it is not impossible that someone will want them taken down and me suspended from this forum. Should that happen my personal email address is lydon_andrew@hotmail.com







5.  RE: National Statistician's statement on the future of consumer price indices

Posted 13-11-2016 09:13

Andrew

John Pullinger was referring to the fact that the ONS promised to submit new evidence for the re-designation of CPIH by September. This it did.

Submitting new evidence, though, is not the same as getting the series approved as a national statistic again. In this case I too think that the decision to make CPIH the headline measure by March is premature.

In particular, Ed Humpherson, the UKSA Director General of Regulation said, referring to the evidence submitted in September, that "we will not base a decision to designate CPIH as a National Statistic on this documentation alone, but will consider whether ONS can demonstrate that it has established the credibility of, and increased users’ confidence in, CPIH." He added that the Regulation arm will consult users.

I do not believe that there is sufficient evidence that ONS has convinced users of the merits of CPIH (and in particular rental equivalence). ONS has put forward arguments for rental equivalence, sure, but, at least as far as the User Group is concerned, I have not seen any evidence yet of the two-way dialogue that is true user engagement. John Pullinger's letter promises further engagement over the coming months. I hope this happens.

------------------------------
Jill



6.  RE: National Statistician's statement on the future of consumer price indices

Posted 13-11-2016 14:39

Jill,

 

I fully agree with you that John Pullinger is correct in that ONS met their formal obligation to submit new evidence on CPIH in September. 

 

Much of this evidence would not appear to be that new as it relies on the deliberations of CPAC which was disbanded following the Smith Review in part as a result of concerns expressed by users. Uncritical reference by ONS to the historical deliberations of CPAC would not seem likely to increase user confidence in CPIH.

 

There also appear to be elements of new evidence missing though in ONS’ defence arguably they have had an excessive amount of work to do.

 

For much of this year UG members, including Andrew, have expressed concerns about the weighting structure of CPIH and that paper will not now be available until December.  It is not clear that this December paper will provide the evidence to answer the questions that have been raised.

 

From a technical statistical perspective there are concerns that the reliance on rental equivalence will produce an unstable index and these need to be fully explored.

 

All the best.

 

Arthur




7.  RE: National Statistician's statement on the future of consumer price indices

Posted 15-11-2016 04:21

I have added a document to the library, "A Better Understanding of the Different Approaches for Measuring OOH Costs". 

The paper suggests a better classification of variants of the three approaches to measuring OOH costs than that provided in one of the papers in the CPIH compendium. It argues that the ONS does not want to recognize the accounting approach as a separate variant of the user cost approach, because it is easier to argue for discontinuing the RPIJ if its OOH series is seen as representing some defective kind of payments index. Not showing the most broadly defined variant of a payments approach as a variant distinct from the one formerly used in the RPI similarly shows a lack of interest in establishing a Household Inflation Index. ONS does not even hint at the different variants of the net acquisitions approach, and the only variant described is not the best one.




8.  RE: National Statistician's statement on the future of consumer price indices

Posted 16-11-2016 12:07

Hi Andrew

I am still reading your paper, but I wonder if you could confirm the meaning to be attached to certain terminology, viz:

House price - this is the price of a house/flat including the land value.  This is how the term is used in the UK.

Dwelling price - this is the price of a house/flat excluding the land value - There is no trading of this nature in the UK, so dwelling prices have to be imputed in some way.

Lot price -  This is the price of a plot of land excluding the buildings on it.  So we should have the equation:

         House price = Dwelling price + Lot price.

Am I correct in the above assumptions?

Best regards

Gareth




9.  RE: National Statistician's statement on the future of consumer price indices

Posted 17-11-2016 15:06
Sorry, I had thought I had replied to the group when I responded to Gareth, but he informs me that I only responded to him. He is right that house refers to dwelling and lot, dwelling to the building or structure only. If this is not clear, one could substitute "dwelling and lot(where appropriate)" for "house". In some cases, e.g. houseboats that may be moored anywhere, house and dwelling will be the same.

For the Canadian new housing price indices, three sets of indices are published: total, house only and land. The breakout of the lot price is provided by the contractors who respond to the New Housing Price Survey. I wouldn't recommend these terms, as the so-called land price relates to the price of a serviced lot, not raw land. Also, although most dictionaries treat "house" as meaning the same as "dwelling" in both conversations and in media reports house prices generally refer to inclusive prices for dwelling and lot together, because a dwelling is usually transacted as a package with the lot it stands on.

It is worth remarking that the common reason given for distinguishing between lots and dwellings, that consumption applies only to dwellings, through depreciation, and not to lots, is not really valid. Anyone with an apple tree on their property that now yields only a fraction of the fruit it used to will not need any convincing on this subject.





10.  RE: National Statistician's statement on the future of consumer price indices

Posted 21-11-2016 11:44

I would like to add support to Andrew's view that the "accounting approach" should be added to the list of general approaches to the measurement of OOH costs.  This approach lies somewhere in between the Payments approach and the User Cost approach.  It is the approach used by the RPI and I believe that the RPI methodology should be included in all comparisons of the percentages changes and weights of OOH costs.  Also I agree with Jill Leyland that the RPI component representing OOH costs should be a published series.

Nonetheless my preference is for the User Cost approach, preferably not based on rental equivalence, but directly estimated as far as possible.  The important advantages of the User Cost approach are that it includes both opportunity cost and depreciation and is responsive to changes in both house prices and interest rates, which most people would deem necessary in order to properly represent the affordability of OOH.  I have suggested to ONS and included on this website a method by which this could be done.  No doubt there are variations on this method possible and these should be researched by ONS.  There is no mention of this in the ONS papers.

Instead the ONS paper refers to Narrow User Cost, which is something it attempted under the old CPAC regime and which used so-called "real" interest rates to offset a perceived capital gain from inflation.  It refers to this only to knock it down immediately.  Instead, ONS should be researching the measurement of true Full User Cost.

The net acquisitions approach (even based on house rather than dwelling prices, which as Andrew argues is superior), seems a very peculiar measure of OOH costs to me.  It does not respond to interest rate changes (not even medium to long term rates) and I don't think the general public will understand why it has ever been considered.

GJ




11.  RE: National Statistician's statement on the future of consumer price indices

Posted 22-11-2016 17:36
Gareth:
Thank you for your support for recognizing the accounting approach as a distinct variant of the user cost approach. I must say, I do find it odd how it is ignored in the literature given that both Canada in its CPI and the UK in its RPI/RPIJ series use this approach.
In my opinion, it is the best approach now used anywhere in official consumer price indices for owner-occupied housing (OOH) for uprating purposes. When you are adjusting pensions or wages for cost-of-living changes, it makes sense to include the actual expenses that homeowners must make in the index, and which generally rise with inflation. If these are replaced by fictive elements such as the imputed rent on owner-occupied dwellings or the opportunity cost of owner's equity, it would make them less suited to their purpose.
The accounting approach does, it is true, include one element that does not relate to monetary transactions, housing depreciation. Take it out and you are left with a limited money payments approach such as the UK RPI had before 1995. To my mind, housing depreciation  is better in than out since owner-occupiers as a class will have to be making expenditures on the housing stock to replace it or else the stock of owner-occupied dwellings and the flow of dwelling services that comes from it will be diminished, and so will their living standards. In the accounting approach, the depreciation component acts like net equity payments in the full payments approach. It also has the advantage that in terms of this model there is a constant stock of owner-occupied dwellings and a constant volume of mortgage debt.
Moreover, it has the advantage that with a few deletions one gets a pretty fair macroprudential inflation measure. This was why the RPIX was the initial choice as the Bank of England's inflation yardstick. The RPI excluding mortgage interest and council tax would have been an even better one, and I am not sure why it wasn't chosen. It's not as good as an HICP with an OOH(NA) component but it should reasonably approximate it.
Just the same, housing depreciation isn't a paid cost, which is why I would support the treatment of OOH in John Astin and Jill Leyland's proposed household inflation index (HII) instead, that would look at actual equity payments. This approach does not guarantee either a fixed stock of owner-occupied dwellings or a fixed volume of mortgage debt. This is probably not such a big concern as it is really a weighting issue rather than a pricing issue, and appropriate normalization to the weights could be introduced if demographic or other factors required it.
 Both the opportunity cost approach you favour and the rental equivalence approach adopted in the CPIH seem to want to define a consumer price series based on consumption as defined in national accounting terms. Although the national accounts uses a rental equivalence approach to measure OOH, the opportunity cost approach is really not so different since one can always choose an interest rate for owner's equity that will make them equivalent. So it seems to me that you are being inconsistent in including transaction costs in your opportunity cost index. In the national accounts, expenses like estate agents' fees are not part of personal expenditure but are a component of residential construction expenditure and hence of gross fixed capital formation. However, if you exclude transaction costs you make it less attractive as a consumer price series for upratings, since these are expenses that all homeowners must pay.
Although I don't like your opportunity cost approach for upratings purposes as much as the broadest payments approach, which is best, or the accounting approach, I would prefer it to the rental equivalence approach. Your user cost approach does give an important role to housing prices, while the rental equivalence approach completely excludes them.
I had meant to write something in defence of the net acquisitions approach, but this e-mail is quite long enough as it is. Suffice it to say that the net acquisitions approach is the most appropriate one for a consumer price series designed as an inflation yardstick for central bank use. It is inappropriate for use in upratings, and for this purpose we both agree that your opportunity cost approach would be more appropriate.
Thank you again for your support for giving the accounting approach status as an alternative to the other user cost approaches.
Best regards,
Andrew





12.  RE: National Statistician's statement on the future of consumer price indices

Posted 23-11-2016 09:30

Dear Andrew and others

I very much like your classification including the Accounting Approach as a separate and distinct way of measuring OOH. As a member of the 1994 RPIAC which supported adding depreciation to the RPI, I think that decision has stood the test of time. The reason the RPI is now considered suspect is nothing to do with its treatment of housing costs - it is more to do with the weighting system, as I understand it.

It would be very simple to transport the RPI OOH treatment into CPI to create a CPIH, and I agree also that stamp duty should be included, particularly when we are now including Council Tax in CPIH. For macroeconomic purposes, a CPIX could be used, excluding the mortgage interest component. This worked well in the UK from 1992 until 2004 when RPI was the main inflation measure.

Best wishes

Andrew Sentance

------------------------------
AndrewSentancePwCSenior Economic Adviser



13.  RE: National Statistician's statement on the future of consumer price indices

Posted 23-11-2016 14:19

Could not possibly agree with Andrew Sentence on the RPI.

From 1997 to 2007 the UK suffered a housing inflation that could be seen from space.

And for most of that period as Andrew himself points out we had the RPI as the target measure.

So I fail to see what can be taken forward from the housing element of the RPI.

------------------------------
AL

Waiting for them to sort out housing in the CPI since 2011

http://www.localisewestmidlands.org.uk/2011/alternative-inflation-report-coping-with-an-inflationary-global-economy/



14.  RE: National Statistician's statement on the future of consumer price indices

Posted 24-11-2016 20:15
Dear Andrew:
All three of us Andrews seem to agree that a macro-prudential inflation measure used as the target inflation indicator of the Bank of England should include house prices and should be properly sensitive to changes in current housing prices. So the appropriate questions would seem to be:
Did the reform of the RPI introduced in 1995 that added a depreciation component make it more sensitive to housing prices?
Was the change, by itself, adequate to have induced a more restrictive monetary policy that would have dampened the housing boom the UK had up to 2007 and the housing bust that ensued afterwards.
 The answer to the first question is obviously yes. The weight of the housing depreciation component in the RPI in 1995 was 30 parts per thousand and it would have been greater in the RPIX series used by the Bank of England as its target inflation indicator.
With regard to the second question, it is worth noting that Gordon Brown announced the UK HICP would be replacing the RPIX as the inflation target of Bank of England in June 2003, although the decision wasn't formally made until December 2003. So the RPIX was effectively the target inflation indicator for only a little more than half of that time. Over the entire period from January 1997 to December 2007, the RPIX series showed a higher or the same inflation rate as the RPIX less depreciation series, i.e. the switch to an accounting approach from a limited payments approach to measuring OOH in the RPI did lead to higher measured inflation during this boom in housing prices. The difference peaked in December 2002 to May 2003, when the difference was 0.8 percentage points. The difference between RPIX and CPI inflation rates peaked in June and July 2003, at 1.7 percentage points.
I would have preferred the RPIX excluding council tax as a target inflation indicator, but this series generally tracks slightly lower than the RPIX itself over the 1997-2007 period.
If you think that UK monetary policy was much too loose over the earlier part of this period, when the RPIX was still the target inflation indicator, perhaps the problem was not so much with the inflation indicator but with the target range. At no time between 1997 and 2003 was the RPIX inflation rate in excess of 3.5%, the old upper bound for the target inflation rate of 2.5%. If the upper bound had been 3.0% on the other hand, the inflation rate would have been at or above that rate in eight different months.
I would not favour calculating an equivalent series to the CPIH with the OOH component based on rental equivalence replaced by one incorporating an OOH component with the same approach as is used in the RPIX, a possibility Andrew Sentance brought up. As a target inflation indicator for the Bank of England it would be inferior to a CPI incorporating an OOH(NA) series, even one with the defective methodology mandated by Eurostat. (I would prefer a gross price-gross weight approach to calculating the new acquisitions series and a gross payments approach to measuring insurance costs.) However, if one had to make a take-it-or-leave-it choice between a CPIH series and Andrew's proposal, I would certainly prefer Andrew's proposal. A target inflation indicator for a central bank that doesn't include house prices is inherently dysfunctional, certainly for a country like the UK.
 
Best regards,

Andrew





15.  RE: National Statistician's statement on the future of consumer price indices

Posted 25-11-2016 10:43

The problem of house price inflation lies in the BOE's instructions from the Treasury, rather than which specific index of general inflation is used.  If the government wants the BOE to control house prices it should set a house price inflation target.  Indeed it could set two targets, one for general inflation and one for house prices.  I think the truth is that the government does not actually want the BOE to do this.  It knows that so long as house building is at such a low level and economic activity so concentrated in the south-east, controlling house price inflation would require very high interest rates which would strangle the rest of the economy.

There is another problem with the remit of the BOE in that subject to keeping inflation low it is also supposed to "support the Government's economic objectives including those for growth and employment".  This is what has given rise to the recent controversial actions by the BOE since the EU referendum - viz. the cutting of interest rates and renewal of quantitative easing at a time when sterling was falling and an interest rate rise would normally be expected.

GJ




16.  RE: National Statistician's statement on the future of consumer price indices

Posted 24-11-2016 17:35
Dear Andrew:

Thank you very much for your message. You and the other people who worked on the 1994 RPIAC report that led to the accounting approach to owner-occupied housing (OOH) should be proud of what you accomplished; it certainly was an improvement that has stood the test of time.  
In his "Statement on future of consumer price inflation statistics in the UK" the National Statistician wrote: "RPIJ shares RPI's other shortcomings, including using a direct measure of house prices to estimate owner occupiers' housing costs." It is hard to know what to make of this terse dismissal of the accounting approach to OOH. If one takes it at face value, he actually seems to be treating it as a defect that the accounting approach relates mostly to paid costs, and not, as with the rental equivalence approach, to fictive costs. Your committee report of 1994 introduced the accounting approach to the RPI. A previous 1974 RPIAC report quite properly rejected the rental equivalence approach which the National Statistician is now so keen on.
The ONS has in its recent publications treated the OOH component of the RPI/RPIJ as a kind of failed payments index. The Bank of Canada has treated the OOH component of the Canadian CPI, also based on the accounting approach, as a kind of failed  opportunity cost index. Presumably the opportunity cost of owner's equity and capital gains were left out of the index because the people responsible were too ignorant or indolent to put them in. Neither agency wants to admit that the accounting approach has its own rationale, different from that of the payments approach or the opportunity cost approach. 
Thank you for agreeing that stamp duty should be in scope in the RPI/RPIJ based on the accounting approach. Everyone seems to agree that stamp duty should be added to the RPIJ. In the Canadian CPI, land transfer taxes are treated as being in scope, although it is, alas, a proxied rather than a priced series. Stamp duty should have been added to the RPIJ no later than March 2015, since the ONS had already started publishing a quarterly OOH(NA) series including stamp duty in December 2014. 
Rather than cancelling publication of the RPIJ in March 2017, the ONS should at that time add stamp duty to it. If we are to get to a household inflation index (HII) along the lines proposed by John Astin and Jill Leyland gradual changes to the RPIJ would seem to be the best way to get there. Some of their proposals, like adding stamp duty or expanding the target population,  are I think, reasonably non-controversial, others are more so. Changing the name of the series from a Household Inflation Index to an Index of Household Payments does not strike me as an innocent change. It rather seems intended to prejudge the debate on whether the accounting approach or a broadly defined payments approach is more appropriate for OOH in a cost-of-living index. (On principle, I would support the latter, but really both series should be calculated in parallel so an informed choice can be made.) The annual Index of Household Payments that Mr. Pullinger promises, with no publication date specified, is an inadequate substitute even for the current RPIJ, let alone the RPIJ that might be.
I have only written about the accounting approach as regards the RPI/RPIJ series and consumer price indices used for upratings, as this reply is already too long. I will write a reply to Andrew Lydon that addresses the use of the accounting approach to OOH in an inflation measure monitored by the Bank of England.
Thank you again for your support, and for being such a consistent champion of having house prices in the Bank of England's target inflation indicator.

Best regards, 

Andrew Baldwin






17.  RE: National Statistician's statement on the future of consumer price indices

Posted 13-11-2016 14:37

Chris,

Thank you for the post.

A couple of questions spring to mind immediately.

First, it is not clear what ONS are proposing and the detail of which weights are being revised for which components would be helpful. 

If a new weight for Council Tax, for example, is included then in principle all the other weights change given that the total is constrained.  The post is not clear on which of the other weights will be changed because of revised source data – national accounts or other – and whether these weight changes will all be back dated until 2005.

Question – which weights will be revised because of revised source data rather than changes solely as a consequence of changes to other weight categories, for example, the inclusion of Council Tax?

A related supplementary question – will the revised historical series also include the change to chain linking methodology that was reported in the article “Assessing the Impact of methodological improvements on the Consumer Prices Index”?

Second, a key issue for the weights analysis raised back in March is the stability of the OOH components – and to a lesser degree all components which use the rental index.  This will require comparisons back to 2005 between the current CPIH index series which excludes Council Tax and that current series – again on the current basis excluding Council Tax – recalculated with the only source data difference being the revised national accounts weights for OOH.  The rent component may also need to be included in the analysis if that historical series was also revised in the national accounts.  It is not clear how these comparisons, or something analogous, will be possible under what ONS are proposing.

Question – how will the analysis of the stability of CPIH and its OOH component be achieved under these ONS proposals?

All the best.

Arthur




18.  RE: National Statistician's statement on the future of consumer price indices

Posted 11-11-2016 14:25
Edited by Shaun Richards 12-11-2016 21:08

Dear James

Thank you for posting the statement from the National Statistician Sir John Pullinger

There is much to discuss but I would like to highlight one point which is housing costs and prices. The statement tells us this.

"RPIJ shares RPI’s other shortcomings, including using a direct measure of house prices to estimate owner occupiers’ housing costs. It therefore is not a good economic measure of consumer inflation."

As an aside this is something of a handbrake turn from an organisation which pushed RPIJ and has therefore led fans of it up their garden path. But if the concept of using house prices is so wrong not only is Eurostat on the wrong path according to Sir John then so must be the IHP he is proposing as if I may quote from Tony Cox's User Group submission on the IHP/HII there is this.

"capital housing costs – should be included. Arguably one of the big failures of our current range of inflation measures is their failure to recognise the increases in housing costs. Although specific housing cost measures do exist. It is important that these are also included in an Index of Household Prices."

Back In March Sir John seemed to be espousing a different path himself as the March version of Stats Life pointed out that the HII had the features below.

"This new consumer price index would include interest payments and the capital cost of houses – items usually excluded from such indices – and would complement a set of macro-economic indices such as the CPIH."

Is out the new in? Or is in the new out?

Regards

Shaun Richards




19.  RE: National Statistician's statement on the future of consumer price indices

Posted 17-11-2016 15:21

All,

Thank you for your responses. Since there have been a number of posts  I will respond to all of the points in one go rather than individually. If you feel I have missed anything do let me know.

As Jill has already noted, the Autumn target was for submitting evidence to the UK Statistics Authority and for providing a further update to Sir Andrew. Both of these have now taken place. Although we are of course looking to get CPIH re-designated as soon as possible the date for re-designation is a matter for the UKSA to decide.

We are confident that by March 2017, when a few further changes such as the inclusion of Council Tax will have been implemented, CPIH will be a robust measure of inflation that users can have confidence in. That is why we have made the decision to go ahead with a March 2017 date for making it our headline measure.

In terms of further user engagement, we have discussed with Tony Cox how we can engage more with the user group, and we would be interested in suggestions from members of this forum about what further engagement they feel should be considered that we have not undertaken to date.

On the use of a direct measure of house prices, ONS and the National Statistician have been consistent in our view that such things should not be included in a consumer price inflation index. The IHP is not, in our view, an index that measures consumer price inflation as the use of a direct measure of house prices means that it is including a significant element of asset prices. As we do not view the IHP as a measure of consumer price inflation, we do not have the same conceptual objections to the use of a direct measure of house prices. The above quote from Stats Life appears to be the Royal Statistical Society’s own words rather than anything said by the National Statistician in his March 2016 letter.

Eurostat has developed an approach to measuring OOH costs that all member states can produce in practice. This rules rental equivalence out, as many countries do not have a sufficiently large rental market, and HICP principles also point to net acquisitions as being the most appropriate measure for the HICP. Whilst the measure should in theory exclude land costs, in practice, member states are unable to disentangle the dwelling costs and land costs, which means that the measure includes an element of asset price. For a UK measure of inflation, however, we do not face the same restrictions, and are free to construct the most appropriate measure for the UK. It is also worth noting that rental equivalence is quite widely used by other countries in their headline inflation measures (see table 2 in the CPIH compendium).

We appreciate that the weights article has taken longer than members of the user group were hoping for. However, we hope you understand that it makes sense to take a bit longer to release new weights information so that we can properly factor in the inclusion of Council Tax.

The article which we are putting together will contain analysis of the impact taken separately (i.e. the impact of revising OOH weights only, and the impact of including Council Tax only), and the impact of both changes implemented together. On the weights topic, including Council Tax means that all the other CPIH weights are reapportioned to constrain to totals. This will be carried back to 2005. The only weight being revised to update with new source data is the OOH component. Neither CPI nor CPIH are being revised for the changes to chain linking methodology and the inclusion of COICOP5.

I hope the above is helpful. I would be happy to answer any further points.                              

Regards

James




20.  RE: National Statistician's statement on the future of consumer price indices

Posted 18-11-2016 20:22

Could I ask why we will only be getting confirmed OOH weights back to 2005, when the national accounts provide a basis for them back to 1997 ?

------------------------------
AL

Waiting for them to sort out housing in the CPI since 2011

http://www.localisewestmidlands.org.uk/2011/alternative-inflation-report-coping-with-an-inflationary-global-economy/



21.  RE: National Statistician's statement on the future of consumer price indices

Posted 20-11-2016 13:20

Money Box, Saturday 19 November,  BBC Radio4

For those who didn't catch this programme it contained a short discussion on CPIH with Jonathan Athow, Director General Economic Statistics, ONS and Andrew Sentance, Senior Economic Advisor to PwC (and a former member of the Bank of England Monetary Policy Committee).

You can 'listen again' to the whole programme at-  http://www.bbc.co.uk/programmes/b082syx2  or register and select this item only further down the page.  If you are listening to the whole programme the CPIH item occurs approximately 14.5 minutes in.

Tony Cox, chair RPI CPI User Group 




22.  RE: National Statistician's statement on the future of consumer price indices

Posted 29-11-2016 10:35

Andrew,

 Thanks for your post.

 The CPIH series runs from January 2005. Therefore, we are producing new weights for the duration of CPIH. We do not have a sufficient sample of rental prices pre-2005 to form an OOH component, hence we will not be producing any earlier weights.

 I hope this clarifies the position.

 Regards,

Chris




23.  RE: National Statistician's statement on the future of consumer price indices

Posted 29-11-2016 13:29

Chris

Does this mean you are putting a significant health warning on the CPI and the RPI before 2005 ?

And how come the data was good enough for the national accounts then ?

------------------------------
AL

Waiting for them to sort out housing in the CPI since 2011

http://www.localisewestmidlands.org.uk/2011/alternative-inflation-report-coping-with-an-inflationary-global-economy/



24.  RE: National Statistician's statement on the future of consumer price indices

Posted 06-12-2016 17:08

Andrew,

Prior to 2005, the CPI and RPI use private rental price data from the local price collection. The data are perfectly adequate for the compilation of a private rental item index but, for constructing a rental equivalence measure of owner occupiers’ housing costs, a larger sample is needed (for more information see the CPIH Compendium sections 2.1 - 2.3).

Prior to 2005, National Accounts expenditure data on imputed rentals are derived from back-casts. For more information on this, see the article Changes to National Accounts: Imputed Rental (Section 4).

Best wishes,

Chris




25.  RE: National Statistician's statement on the future of consumer price indices

Posted 09-12-2016 10:19

Chris

The links you sent me don't seem to include anything that I have not seen from the ONS about rental equivalence that I have not seen before, or do you think they do.

Could I therefore ask what the difference is between the rental price indice figures for 2004 and 2005 are in terms of their composition ?

------------------------------
AL

how the US CPI actually worked to contain housing inflation

http://www.localisewestmidlands.org.uk/2012/housing-reflation-and-housing-inflation-alternative-inflation/



26.  RE: National Statistician's statement on the future of consumer price indices

Posted 14-12-2016 15:20

Andrew,

After 2005, private rental indices are based on a very large sample of rental prices – initially around 200,000, but rising to around half a million since 2011 - collected by VOA, Welsh and Scottish Governments. Indices are created for each region, property type and furnished status. These indices are then weighted together, using stratum weights that reflect the private rental market, to form an index for private rental prices. These same indices are also weighted together, using stratum weights that reflect the OOH market, to form the rental equivalence OOH index.

Prior to 2005 the item index for private rents is based on a much smaller sample of rental prices - around 1,200 - collected as part of the ongoing price collection for CPI. Indices are compiled based on region and furnished status, but the property type split is not available. The small sample sizes also mean that the stratum indices can be particularly volatile.

Kind regards,

Chris




27.  RE: National Statistician's statement on the future of consumer price indices

Posted 14-12-2016 20:13

Chris

As we now have the series of OOH valuations going back to 1985, could you point to instances of volatility that would preclude their use for weighting the CPI, but still allows them to pass fit for the Blue Book.

------------------------------
AL

how the US CPI actually worked to contain housing inflation

http://www.localisewestmidlands.org.uk/2012/housing-reflation-and-housing-inflation-alternative-inflation/



28.  RE: National Statistician's statement on the future of consumer price indices

Posted 06-01-2017 17:06

Andrew,

The volatility is in the price series prior to 2005. This is distinct from the National Accounts expenditure data that we use to weight the CPIH, which are not volatile.

I hope this clarifies the point.

Regards,

Chris




29.  RE: National Statistician's statement on the future of consumer price indices

Posted 07-01-2017 09:38

To be absolutely clear Chris are you saying you would be confident of the National Accounts data for weights prior to 2005, but are not confident of the series data that would be given that weight in the overall CPI-H ?






30.  RE: National Statistician's statement on the future of consumer price indices

Posted 26-01-2017 16:44

Andrew,

Yes, that is correct. The National Accounts data are the best source of information for weights before (and after) 2005. The rental price series, however, is volatile prior to 2005.

Regards,

Chris




31.  RE: National Statistician's statement on the future of consumer price indices

Posted 27-01-2017 09:32

Chris


I did a post back in the summer to outline my reaction to the  2016 Blue Book revisions.


In that latest Blue Book Owner-occupied housing has a bigger proportion of household consumption in 1996 before the great UK house price boom began than it did at the end of the house price boom, or even most recently.


This was not the case with the figures in the 2015 Blue Book. At least the fact that the value of Owner-Occupied Housing increased as a proportion of household consumption since 1996, gave the 2015 figures a plausibility that the current Blue Book does not have. And this Blue Book is as you acknowledge the source of your weights from 2006 onwards.


I am not aware of any comparisons of house prices to incomes during that boom period that does not indicate that prices doubled in relation to incomes. And even if you look at it in terms of rental value having nothing to do with purchase prices, we would not have an issue with Generation Rent if the value of owner-occupation was falling in relation to household consumption. Indeed the younger generations would be happy with being Generation Rent.


Is there not a plausibility problem here ? Or am I missing something ?






33.  RE: National Statistician's statement on the future of consumer price indices

Posted 29-01-2017 22:55
Edited by Arthur Barnett 30-01-2017 13:42
John,

National accounts produce both volume and current price measures.  The latter are listed after the volume measures in - Gross Domestic Product (GDP) - Office for National Statistics

I will admit though that I don't know whether ONS use the volume or current price measures to calculate the weights.  I've tended to assume that because consumer price indices are indexes of price relatives rather than prices that it is the current price measures that are used but I may be wrong.

Mind you total GDP at £471 billion for the chain linked volume measure is not that much different from the £487 billion current price measure.

Perhaps ONS could confirm which measures are used.

(Note this is a later edit - John's post Is no longer on SUN but the question to ONS about whether a volume or current price measure is used still remains relevant and an answer would be much appreciated.)

Arthur


34.  RE: National Statistician's statement on the future of consumer price indices

Posted 30-01-2017 22:19

Chris,

 

I have been looking at the various pieces of data I have available to see if there is anything that might throw some light on Andrew Lydon’s point about the reduction in the OOH weight over time.  The data does appear to support the need to explain the historic reduction in OOH weights.

 

I have used 3 tables which are given below.  These tables provide the different versions of the OOH weights, DCLG dwelling stock data and private rental weights that Richard Campbell provided a couple of years ago.

 

The analysis is constrained to 2005 to 2014 because of the limitations of the DCLG data.

 

The main assumption is that the weight is a value weight and therefore can be approximated as quantity multiplied by price – quantity being DCLG OOH dwelling stock. 

 

The latest revised CPIH OOH weight falls by around 10% between 2005 (19.6%) and 2014 (17.4%).  The DCLG owner occupied dwelling stock fell by only about 2% between 2005 (18,130 thousand) and 2014 (17,712 thousand).  Therefore the reduction in the dwelling stock quantity is far less than the reduction in the OOH weight implying a larger reduction elsewhere in the weight calculations.  The rental price index increased by about 25% between 2005 and 2014 so price cannot be the source of this reduction.

 

Assuming that the dwelling stock fall (-2%) and the rental price index increase (+25%) are the only factors affecting the OOH weight – this is clearly not the case as the weight change is affected by DCLG data lags, Blue Book methodology and weight changes elsewhere – then this would imply a 22% increase between 2005 and 2014.  This simple assumption would suggest a 2014 weight much higher than the current 17.4% at about 24%.

 

The above is not to suggest the “right” weight is 24% but to provide support to Andrew’s query about the credibility of the reduction in OOH weights. 

 

I may have got my logic wrong – in which case profound apologies – but there does seem to be the need for an explanation of how the national accounts methodology results in reducing historical OOH weights.

 

In contrast to the latest revised OOH weights the same logic gives very different answers for private rents and the first OOH weights – neither result in an implied negative.  The private rents calculation comes up with an increase which is close to the 25% increase in the rental price index while the first OOH weights calculation gives an increase twice as high as the increase in the rental price index.  These last two examples are illustrative as neither of these weights may be calculated using the DCLG dwelling stock figures.

 

All the best.

 

Arthur.

 

 

Note: I don’t think that the recent issue about whether the OOH weights are based on volume or current price national accounts estimates materially affects the above.  If the former are used then it gives the implicit price a longer lag and makes it more difficult to identify.

 

Table A: posted in a response to Tanya a fortnight ago.

 

 Table B: DCLG table 101 extract

 

 

Table C: Private rents plus OOH (1st revise) weights table from Richard Campbell

 

 

Please note that the private rents price series is based on the old data source prior to 2013 for consistency with the CPI. The Imputed rents (aka OOH) are mostly based on the new data in recent years (approx 3.6 pp 1000 cover Northern Ireland OOH where the old private rental data are still used). As you go back towards 2005, a higher proportion of the OOH price series will be based on the old private rental data source prior the new sources of data for Welsh and Scottish data coming on stream. At most, this will amount to around a further 23 pp 1000).




35.  RE: National Statistician's statement on the future of consumer price indices

Posted 07-02-2017 09:32

Arthur, Andrew,

The reason that the OOH weight has been falling is because imputed rentals expenditure is increasing less quickly than expenditure for other items in the basket.

From Blue Book 2016 Table 6.4, between 2005 and 2015 expenditure on imputed rentals increased from 148028 to 190140 (and this was a year on year increase). This is an increase of approximately 28%. However, expenditure on food and non-alcoholic beverages has increased by 40% over the period, clothing by 43%, actual rentals by 87%, and transport services by 68%. Because the weights are constrained to sum to 1000, and because expenditure on other items has increased more, the expenditure share for imputed rentals is less in 2015 than it was in 2005. (Note that this is slightly different data than we’d use to calculate weights in practice, as this is a BB16 consistent time series rather than price updated latest Blue Book expenditure; however, the point still stands).

We do not consider this to be implausible because owner occupation has been falling over the period. It does not seem unreasonable that, although prices may be rising, as far as expenditure is concerned (price x quantity), falling owner occupation would lead to expenditure shares increasing at a slower rate than other items in the CPIH basket.

I can confirm that the national accounts data used to produce expenditure shares are taken from the current price series.

Kind regards,

Chris




36.  RE: National Statistician's statement on the future of consumer price indices

Posted 07-02-2017 09:45

So you are saying that OOH consumption value has not risen as much as food and non-alcoholic beverages ?


Is that remotely plausible when we have Generation Rent and house prices are double what they were in relation to incomes in 1996 ? In many instances food prices have actually fallen in nominal terms in recent years.


Have the rental prices of the sort of properties that acrue actual rentals not been the sort of residential properties that have an imputed rent ascribed to them ?  The former is 3 times the later.


AL










37.  RE: National Statistician's statement on the future of consumer price indices

Posted 22-02-2017 10:17

Andrew,

Expenditure is a function of both price and quantity. The fact that house prices have increased is not enough on its own to drive an increase in expenditure. If owner occupation has fallen this will have a downward pull on expenditure. The fact that expenditure has increased less quickly than for other categories is consistent with the ‘Generation Rent’ idea and falling owner occupation. Similarly, if food prices have fallen, and consumers have capitalised on this by purchasing more of that food, then it is entirely possible that expenditure may increase more steeply.

Kind regards,
Chris




38.  RE: National Statistician's statement on the future of consumer price indices

Posted 22-02-2017 10:29
Chris

We should not need to speculate on this it should be demonstrable.

How far has owners occupation fell by comparison to rental value ? I must say my impression  is prices have risen more dramatically than the current fall off in owners occupation.

But in any event this is a national accounts matter, and should not be distorting a CPI. The Generation Rent issue should be showing up in a proper inflation index. And it would show up in an index constructed like that in the USA.

------------------------------
AL

how the US CPI actually worked to contain housing inflation

http://www.localisewestmidlands.org.uk/2012/housing-reflation-and-housing-inflation-alternative-inflation/
------------------------------



39.  RE: National Statistician's statement on the future of consumer price indices

Posted 07-02-2017 22:14

Chris,

 

Thank you for your post – most useful.

 

What it appears to imply is that in the context of price indices the RE concept, or at least the implementation of the concept by ONS, may be more of an issue than national account methodology.

 

Unless ONS have adjusted the rents to remove landlords’ capital costs – interest etc – then RE as implemented by ONS includes proxy capital costs.  These capital costs will be different from those that apply to owner occupiers because of, for example, issues like tax relief.

 

ONS may argue that RE excludes interest costs etc but arguably ONS have simply substituted landlords’ costs for owner occupiers’ costs.  I suspect that this may be difficult for owner occupiers to find credible.

 

Thank you for confirming that the weights use the national accounts current price series.

 

And to confirm that the questions in my previous post which largely relate to historical revisions and future estimates remain important though are not particularly urgent.

 

All the best.

 

Arthur.




40.  RE: National Statistician's statement on the future of consumer price indices

Posted 10-01-2017 09:46

Could I ask again for some clarity about what is in the documentation that was released in December about OOH and Rental Equivalence that was not presented before ?






41.  RE: National Statistician's statement on the future of consumer price indices

Posted 26-01-2017 16:51

Andrew,

Thank you for your query. I’d be happy to clarify what information has and hasn’t been presented before.

The CPIH Compendium brings together information which we have previously published, with some new sections (mainly in section 3). The Compendium aims to provide users with one document which provides a clear explanation of the rationale for our choice of the rental equivalence approach to measure owner occupiers’ housing costs (OOH), how it is constructed and any potential issues with quality. We also updated an article detailing the Users and Uses of consumer price statistics.

The following articles, however, are entirely new:

I hope that this clarifies the publications.

Kind regards,

Chris




42.  RE: National Statistician's statement on the future of consumer price indices

Posted 20-11-2016 21:07

James,

 

Thank you for the helpful information about the weights paper.

 

I will admit to being a little surprised that CPI and CPIH are not at least being considered for being revised historically to include the changes to chain linking methodology and the inclusion of COICOP5.

 

I read Chris’ paper “Assessing the Impact of methodological improvements on the Consumer Prices Index” some time ago and intended to post a reply but as you are only too well aware there have been a lot of things to do of late with all the publications over CPIH re-designation.

 

My surprise arises because, although Chris’ paper does not explicitly explore the cumulative long term effects of the changes, his analysis hints that the cumulative effect may be approaching the scale of the earlier changes to CPIH that were in part responsible for its loss of National Statistic status. 

 

Apart from the lack of analysis of cumulative effects the paper is problematical and confusing in that, for example, the commentary about the CPIH on page 9 states Figure 7 refers to CPIH whereas the title of that figure refers to CPI.  Sorting out these drafting difficulties is not though the major requirement.  What is required is an analysis of the cumulative long term effect of the changes which is currently missing from Chris’ paper.

 

As a start to get a feel for the potential scale of any problems that might become apparent I would suggest a couple of small tables comprising 6 to 12 numbers in total.  Chris will have these numbers to hand so there will be little work involved.

 

Could ONS provide a table for each of CPI and CPIH of the index values at the beginning and end of the period in Chris’ analysis with the following breakdowns –

 

·         Current series

·         Inclusion of COICOP5 effect only

·         Inclusion of chain linking effect only

·         Inclusion of both effects.

 

Thank you again for your very helpful post.

 

All the best.

 

Arthur




43.  RE: National Statistician's statement on the future of consumer price indices

Posted 18-01-2017 09:43

Arthur,

We have uploaded the tables you requested to the ad-hoc release section of the ONS website here. The tables contain estimated CPI and CPIH index values for 2005 and 2014, taking the impact of COICOP5 and the double price update separately and together. I hope that this gives you the information you need, but do get in touch if you need anything more.

Regards,

Chris




44.  RE: National Statistician's statement on the future of consumer price indices

Posted 07-12-2016 09:16
  |   view attached

Please see the attached User Group response to the National Statistician's statement on the future of consumer price indices of 10 November 2016.

Tony Cox

chair, RPI CPI User Group




45.  RE: National Statistician's statement on the future of consumer price indices

Posted 07-12-2016 09:25

One wonders whether the Tripartite Committee of the Treasury, Bank of England and ONS is still operating in the background.

There is a huge conflict of interest here that I will mention elsewhere.

------------------------------
AL

Waiting for them to sort out housing in the CPI since 2011

http://www.localisewestmidlands.org.uk/2011/alternative-inflation-report-coping-with-an-inflationary-global-economy/



46.  RE: National Statistician's statement on the future of consumer price indices

Posted 08-12-2016 03:49
Dear Tony:
Thank you very much for the User Group response that you have just sent out.
I appreciate the defence of the continuation of the RPIJ that it provides. As you say, the National Statistician's letter isn't clear what he finds wrong with the accounting approach to owner-occupied housing (OOH) used in the RPI/RPIJ. Mr. Pullinger wrote that he didn't like it because it was "using a direct measure of house prices to estimate owner occupiers' housing costs". As you pointed out, this was in the context of calculating an index for depreciation. So is the problem that a housing price index was being used where a dwelling index was more appropriate? In the Canadian CPI, which also uses an accounting approach to OOH, a dwelling index is used, and if this was really a sticking point then one could surely substitute a dwelling price index for the house price index in the RPI/RPIJ calculation. Or does the National Statistician believe that an appropriate OOH series should not include housing prices at all, like the rental equivalence approach? It is hard to see that this makes much sense.
As I have said before the difference between the accounting approach to OOH and the broadly defined payments approach to OOH proposed for an HII are small. If you ignore the exclusion of stamp duty from the RPI/RPIJ, which naturally should be included, along with other transaction costs, there are only two:
The housing depreciation component would be replaced by a component for net equity payments, i.e. downpayments plus the equity portion of mortgage payments less sales of homes;
Renovations expenditures would be added to the OOH component.
These are not huge differences, and reasonable people could dispute whether even the second change is appropriate. (I think it is.) This makes me believe that the most appropriate way to create an HII would be by gradual modifications of the RPI/RPIJ.
 Thank you again for sending us the User group response.
Best regards,
Andrew





47.  RE: National Statistician's statement on the future of consumer price indices

Posted 08-12-2016 17:03

I'd just like to add my support to the User Group's response. I, too, am disappointed that ONS appeared to have ignored well-argued comments from users and other interested parties without addressing our concerns directly, merely having recourse to unjustified statements about what is best and spurious arguments regarding consumption. It may be worth pointing out to ONS staff that RPI, CPI and CPIH are prices indices (the clue is in the name) and are intended to measure changes in prices, not changes in consumption. The measurement of changes in consumption is the job of the quantity indices of National Accounts.




48.  RE: National Statistician's statement on the future of consumer price indices