One moment, processing...

Printer-friendly version

RPICPI User Group

search criteria = ALL
RPICPI User Group sorted by thread
 
  National Statistician's announ...
January 10, 2013 2:38 AMBethan Evans
  RE:National Statistician's ann...
January 10, 2013 3:00 AMTony Cox
  RE:National Statistician's ann...
January 10, 2013 7:29 AMGareth Jones
  RE:National Statistician's ann...
January 10, 2013 6:47 AMAndrew Lydon
  RE:National Statistician's ann...
January 10, 2013 9:16 AMShaun Richards
  RE:National Statistician's ann...
January 10, 2013 9:58 AMJill Leyland
  RE:National Statistician's ann...
January 23, 2013 12:44 PMArthur Barnett
  RE:National Statistician's ann...
January 24, 2013 2:24 AMJill Leyland
  RE:National Statistician's ann...
January 24, 2013 6:15 PMShaun Richards
  RE:National Statistician's ann...
January 25, 2013 10:14 AMJohn Wood
  RE:National Statistician's ann...
January 25, 2013 2:38 PMGareth Jones
 

1.
National Statistician's announcement on the RPI consultation
From: Bethan Evans
To: RPICPI User Group
Posted: January 10, 2013 2:38 AM
Subject: National Statistician's announcement on the RPI consultation
Message:

Dear all,

The National Statistician, Jil Matheson, has published her recommendations following the consultation on options for improving the Retail Prices Index (as of 7:00am, 10 January 2013).

The formal announcement can be found in the following link: http://www.ons.gov.uk/ons/rel/mro/news-release/rpirecommendations/rpinewsrelease.html

and a link to the CPAC papers for more information: http://www.ons.gov.uk/ons/guide-method/development-programmes/other-development-work/consumer-prices-advisory-committee/cpac-papers/index.html

The National Statistician has concluded that the formula used to produce the RPI does not meet international standards and recommended that a new index be published. A new RPI-based index will be published from March 2013 using a geometric function (Jevons), to be known as RPIJ. In developing her recommendations Jil Matheson also noted that there is significant value to users in maintaining the continuity of the existing RPI's long time series without major change, so that it may continue to be used for long-term indexation and for the index-linked gilts and bonds in accordance with user expectations

The National Statistician has also recommended that improvements to the measurement of private housing rents from using an alternative data source should be implemented in the February 2013 RPI indices (published on 19 March 2013). As this change affects the RPI it is subject to the view of the Bank of England and, if necessary, approval by the Chancellor of the Exchequer.

Kind regards,

Bethan Evans
Prices Division
Office for National Statistics

Be the first person to recommend this.


2.
RE:National Statistician's announcement on the RPI consultation
From: Tony Cox
To: RPICPI User Group
Posted: January 10, 2013 3:00 AM
Subject: RE:National Statistician's announcement on the RPI consultation
Message:


Thanks to Bethan for posting the the National Statistician's recommendation, following the ONS consultation on - "options for improving the Retail Prices Index".


The National Statistician has effectively recommended option 1 - no change to the RPI. However a new index RPIJ will be introduced from March 2013 which may go further than the change proposed in option 4 of the consultation. This is because it is apparently proposed to rely purely on the use of the Jevons formula, eliminating the small use of the Dutot used in the CPI.


As you know we responded to this consultation and were particularly concerned at the lack of empirical evidence supporting change. (Our full response is in the library section, 29 November 2012.)


We welcome the decision not to change the RPI at this time, but consider the ONS approach was flawed. The consultation was premature. Research into the size and causes of the differences between the RPI and the CPI was already under way. It should have been allowed to conclude. We would expect decisions to change an index for which billions of pounds of investments are linked to be taken in a considered manner based on comprehensive evidence. This is especially so as any change could affect nearly everyone in the UK - through the pensions they get or hope to get, or the prices they pay for RPI-linked services. Moreover, whilst the ONS pointed to "price-bounce bias" as a reason to change the RPI, no attempt has been made to measure its effect in practice.  Many users were left to question whether one of the key rationales for the consultation did indeed warrant such a fundamental change or instead was an abstract theoretical concern.

 

The consultation was also unbalanced as it concerned itself with only one of the formula used to compile price indexes - the Carli formula used in the RPI. The impact of the different formulae used in both the RPI and CPI depends on their interaction with price movements, price collection approaches and other factors which should have been fully investigated and quantified before any proposal for change was considered.


We plan to hold a meeting in the near future to discuss how the User Group should respond to this recommendation in particular, but also our work programme for the coming year. Details will appear on this web-site soon. In the meantime please let us know your reaction to this announcement and any ideas you have for areas on which the User Group should focus, by posting your thoughts to this web-site.


Tony Cox

Chair








Show Original Message
Be the first person to recommend this.


3.
RE:National Statistician's announcement on the RPI consultation
From: Gareth Jones
To: RPICPI User Group
Posted: January 10, 2013 7:29 AM
Subject: RE:National Statistician's announcement on the RPI consultation
Message:
I think in future that people will focus on CPIH not RPIJ.  RPIJ effectively already exists (by adjusting the RPI for the published formula effect, roughly speaking).

Our effort should therefore be to modify CPIH so that it produces a realistic and accurate measure of inflation.  Fortunately CPIH is not constrained by either UK or EU legislation.

A few lines of investigation suggest themselves:

a) improvements to the RE methodology for OOH costs (and also private rents data)
b) more data to allow weighting and stratification at the lowest level of the index so as to minimise the need to use a formula.
c) more objective criteria to decide which formula to use for which products (assuming no weights available)
d) attention to the treatment of outliers, whatever formula is used (also consideration of the median of price relatives)
e) decisions on the deliberate omission of certain products from the index because of their peculiar price behaviour (e.g. fashion goods).
f) quantification of all effects used in argument - notably "price bounce".
g) sampling and data collection procedures and the monitoring of them.

GJ








Show Original Message
Be the first person to recommend this.


4.
RE:National Statistician's announcement on the RPI consultation
From: Andrew Lydon
To: RPICPI User Group
Posted: January 10, 2013 6:47 AM
Subject: RE:National Statistician's announcement on the RPI consultation
Message:

Having read the announcement and compared it to what was said on the BBC this morning, it all looks like a another cock-up to me.

At first it seemed that they recognised the major flaws in the RPI, and were going to design a new index (up to international standards). Well, you can't argue with that.

However, from the announcement text it seems this will only be another fiddled about with RPI ( or is it an RPIJ). This will be launched in the Spring along with the CPIH.

So this Spring we will have an RPI, RPI-X, RPIJ. CPI, CPIH. I am looking forward to seeing how they will spin any clarity out of all this. 

Then they are going to have to rely on the BBC etc to convey this.

If I have got any of this wrong, I would welcome being corrected.

AL





Show Original Message
Be the first person to recommend this.


5.
RE:National Statistician's announcement on the RPI consultation
From: Shaun Richards
To: RPICPI User Group
Posted: January 10, 2013 9:16 AM
Subject: RE:National Statistician's announcement on the RPI consultation
Message:
Dear Bethan

Thank you for letting us know. This is  a victory for those of us here who voted for the no change option. According to the response which Tony Cox kindly posted I believe that it was 80% of those who voted.

From the point of view of the ONS and the National Statistician this has become something of a shambles. We now have inflation in the number of inflation measures! Still at least many news organisations like the BBC read out your press release rather than looking further so the full extent of what has taken place here will not be explained to many.

For those who have followed the debate I have posted my views here.

http://www.mindfulmoney.co.uk/wp/shaun-richards/what-is-happening-in-spains-economy-and-what-is-the-outlook-for-2013/

And before I finish I would like to say thank you to everyone who contributed to the debate.

Regards

Shaun Richards 






Show Original Message
Be the first person to recommend this.


6.
RE:National Statistician's announcement on the RPI consultation
From: Jill Leyland
To: RPICPI User Group
Posted: January 10, 2013 9:58 AM
Subject: RE:National Statistician's announcement on the RPI consultation
Message:



Here is the RSS statement in response to this morning's announcement:

Royal Statistical Society statement in response to the announcement of the National Statistician on the outcome of the consultation on the Retail Price Index

 

10 January 2013

 

 

The Royal Statistical Society welcomes the decision of the National Statistician and the UK Statistics Authority to continue publication of the RPI unchanged while continuing the ONS research programme in consultation with users.

 

How inflation measures are calculated is not just a technical issue for professional users but one of widespread public importance affecting millions of UK citizens.

 

The ONS instituted a comprehensive research programme into the formula effect (that part of the difference between CPI and RPI inflation rates which is due to the use of different index formulae) in 2011. This was rightly focused primarily on clothing which is a major contributor to the effect. The formula effect generally, and particularly in the case of clothing, is dependent not just on the choice of index but also on the characteristics of the price movements and levels being measured, sample design, choice of base period and price collection methods. The research should therefore continue to be comprehensive, include all aspects of methodology and cover the CPI as well as the RPI.  Simply replacing the Carli formula with the Jevons (geometric mean), as will happen in RPI-J, is not optimal.

 

The RSS also welcomes the UK Statistics Authority decision to review the governance arrangements and structures supporting the production of price indices to ensure that these statistics best meet user needs in the future.

 

Importantly, we agree with others affected that this announcement should be a first step in the creation of a coherent family of indices meeting the needs of different user communities. A particular need is to consider the different requirements of an index used for uprating purposes with that used for macroeconomic needs.







Show Original Message
Be the first person to recommend this.


7.
RE:National Statistician's announcement on the RPI consultation
From: Arthur Barnett
To: RPICPI User Group
Posted: January 23, 2013 12:44 PM
Subject: RE:National Statistician's announcement on the RPI consultation
Message:

Members of the group may be aware of the changes that ONS have made to long term real GDP due to revisions to the GDP deflator as set out in this blog by Sam Williamson, Research Professor at University of Illinois, Chicago, and President of Measuring Worth -

 http://www.econbrowser.com/archives/2013/01/guest_contribut_30.html

I have only recently seen the blog and have not attempted to check the detail of the analysis.

If correct the analysis provides another perspective on the importance of achieving consensus on changes to price index methodology.

 Arthur Barnett


Show Original Message
Be the first person to recommend this.


8.
RE:National Statistician's announcement on the RPI consultation
From: Jill Leyland
To: RPICPI User Group
Posted: January 24, 2013 2:24 AM
Subject: RE:National Statistician's announcement on the RPI consultation
Message:

Hi Arthur and all - just to say that the RSS (specifically a member of its National Statistics Advisory Group) noticed this blog too. We have raised the issue with the ONS and I will ask that what we learn is posted to the RPI CPI forum when we have a reply.





Show Original Message
Be the first person to recommend this.


9.
RE:National Statistician's announcement on the RPI consultation
From: Shaun Richards
To: RPICPI User Group
Posted: January 24, 2013 6:15 PM
Subject: RE:National Statistician's announcement on the RPI consultation
Message:


Hi Arthur and Jill

Thanks for your thoughts on this fascinating article and development. When I covered the replacement of RPI numbers with CPI in the GDP numbers I was thinking forwards and hadnt considered fully the possibility that history may get a rewrite.

I have done a spot check on the data for 1997 as a test and it appears to back up the arguments raised. So I await any updates.

Regards

Shaun Richards
http://www.mindfulmoney.co.uk/wp/author/shaun-richards/




Show Original Message
Be the first person to recommend this.


10.
RE:National Statistician's announcement on the RPI consultation
From: John Wood
To: RPICPI User Group
Posted: January 25, 2013 10:14 AM
Subject: RE:National Statistician's announcement on the RPI consultation
Message:


Before getting into a funk about the rewriting of history (a common practice in National Accounts), it's important to note an essential difference between the GDP deflator and the RPI/CPI price indices. The latter are Lowe indices (base-weighted arithmetic means) but the GDP deflator is a derived Paasche index (current-weighted harmonic mean), a necessary consequence of the fact that GDP itself is a (chained) Laspeyres index. To ensure the (nearly) correct use of Paasche deflators in GDP, National Accounts use RPI/CPI indices at as low a level of detail as possible, very often at the elementary aggregate level. However, even at that level, the use of arithmetic means in these elementary aggregates creates a bias relative to the correct Paasche formula. The Jevons formula used in the CPI elementary aggregates is likely to produce a better approximation to the desired Paasche formula than the Carli formula used in RPI. This is the justification for the change, which addresses a well-known and long-standing anomaly in the deflation of National Accounts. It is legitimate to question the accuracy of the change, particularly as it involves an extensive, backward extrapolation of CPI elementary aggregates, but the reason for making it is sound and in the correct direction: Paasche indices usually show lower inflation rates than Lowe indices, hence the effect of the change is to reduce the rate of increase in the GDP deflator and increase GDP growth. Because the GDP deflator has an implicit Paasche formula, very different from those discussed for RPI/CPI elementary aggregates, this change to GDP measurement has little relevance to the question as to what is the correct formula for these elementary aggregates.



Show Original Message
Be the first person to recommend this.


11.
RE:National Statistician's announcement on the RPI consultation
From: Gareth Jones
To: RPICPI User Group
Posted: January 25, 2013 2:38 PM
Subject: RE:National Statistician's announcement on the RPI consultation
Message:
It would be interesting to see the GDP revisions broken down by industry subsectors.  Presumably the main revisions will all be in subsectors of Retail.

GJ 






Show Original Message
Be the first person to recommend this.