|
|
|
From:
|
Gareth Jones
|
|
To:
|
RPICPI User Group
|
|
Posted:
|
February 27, 2013 6:11 AM
|
|
Subject:
|
RE:National Statistician's response to consultation on options for improving the RPI
|
|
Message:
|
|
|
A further thought on this matter relates to biases based on an assumption or premise(e.g. a distributional form). Quantification should be relative to actual data not an assumption. One could of course test the assumption first, but such tests are rarely conclusive. It is always better to work from a real life dataset. GJ
Show Original Message
-------------------------------------------
Original Message:
Sent: 27-02-2013 06:03
From: Gareth Jones
Subject: National Statistician's response to consultation on options for improving the RPI
I have been looking at some of the expert advice ONS has received on the RPI consultation responses and at related material. There is much reference to possible biases of various kinds with different indices or formulae. However it is all theoretical and unquantified.
In Statistics generally, we live with biases of many kinds because the biases are small enough not to have a significant practical effect, or because there is nothing better available. Some things referred to as bias are actually more like variance, because their direction can change.
I think the whole RPI/CPI debate would benefit from quantification of any biases or other effects brought into the argument. I suspect many theoretically possible effects are trivial in practice and could be ignored.
Such quantification needs to be done using real life UK price data. At the moment this means that ONS has to do the work, but I think consideration should be given to releasing the datasets for research by others, which should speed up such work. The datsets are not confidential, being publicly quoted prices. There may be computer security issues if researchers are using ONS computers for the research, but arrangements have been put in place for similar research on other datasets, so there is no real problem here.
GJ
-------------------------------------------
Original Message:
Sent: 22-02-2013 07:13
From: Bethan Evans
Subject: National Statistician's response to consultation on options for improving the RPI
Dear All, I would like to highlight that this morning the response to the National Statistician's consultation on options for improving the RPI was published on the ONS website. Alongside the response ONS has published a letter from the Bank of England regarding the RPI changes, advice provided by independent experts and a copy of responses to the consultation. Links to the releases are available on this page. You may also be interested to know that ONS will be publishing articles explaining the new CPIH and RPIJ measures of inflation on 12 March. A back series will also be published in addition to describing the new measures. Links to the articles will appear on this page.
Kind regards, Bethan
Prices Division
Office for National Statistics
|
|
Be the first person to recommend this.
|