The Economic Affairs Committee launches its short inquiry into the use of RPI with two evidence sessions. Witnesses will be asked if we should stop using the Retail Price Index (RPI) as a measure of inflation, what reasons are there for keeping it, and what impact would changing RPI have on the people and organisations who use it.
Following comments to the Committee by the Governor of the Bank of England in January 2018 that it may be time to transition away from using the retail price index (RPI), as referred to in the Committee's most recent report, Treating Students Fairly: The Economics of Post-School Education, the Economic Affairs Committee of the House of Lords is conducting a short inquiry into the use of RPI. It will conclude before the summer recess.
Tuesday 12 June in Committee Room 1, Palace of WestminsterAt 3.35pm
At 4.35 pm
What are the differences between how RPI and CPI are calculated?
What are the arguments against scrapping RPI altogether?
How much could the Government save in reduced debt interest payments if it switched index-linked gilts onto CPI?
Do you agree with the Chancellor of the Exchequer that moving from RPI to CPI will have a physical cost and will shift the burden of taxation between different groups of taxpayers? Who will be the winners, and who will lose out?
Why do you think RPI is "a poor measure of inflation"?
In the ONS's March 2018 paper on RPI, the National Statistician said that a reason for not changing RPI was that some users valued the continuity of the index, despite its flaws. Can you explain why you think it is more important for a statistic to be consistent than correct
Can you explain the process for making changes to RPI?
What would you like to see happen to RPI?
Staff to the Committee will be in attendance this evening, and we have emailed the details to the members: the unfortunate short notice and the busy parliamentary schedule currently means it may be unlikely for them to attend. We will report back to them on the event nevertheless."
I look forwards to meeting them later.RegardsShaunhttps://notayesmanseconomics.wordpress.com/
David, I agree with you about the witnesses on January 12, who were uniformly poor. The witnesses the week after were much better, particularly Simon Briscoe. They all seemed to agree that there was a need for at least two consumer price series, a household-oriented measure and a macroeconomic measure. The first four all seemed to be looking for one index to rule them all and in the darkness bind them. For example David Johnson said: "…there is no particular reason for using RPI and CPI in different cases. There is no argument there." I don't think Dr. Courtney was the witness I would most have liked to have seen instead of the first four, although he certainly would have provided a corrective to the absolute dismissal of the use of the Carli formula by the witnesses. I was particularly disappointed in Chris Giles as he mentioned the Dutot formula in his testimony but seemed to have no problem with it. I suspect it is still too much used in the UK consumer price series, even though it can generate noisy series as compared to the Jevons particularly for luxury goods. He seemed to understand that well enough when he was a member of the CPIAC. Now he seems to have forgotten all about it.
The HCIs didn't come up at all on January 12, as they surely would have if John Astin, Jill Leyland, or Helen Sands had been called on to testify. Shaun Richards would surely have pointed out that the case for moving from RPIX to the UK HICP as the Bank of England's target inflation indicator was weak, and that it implicitly meant an increase in the target rate of inflation. Also, the Lords would not have been misinformed about the treatment of owner-occupied housing in the RPI, if I had been a witness. It was absurd to say that the accounting approach to OOH used in the RPI was some kind of muddled mix of the payments and user cost approaches. It is a separate variant of the user cost approach, distinct from the opportunity cost approach. It was disappointing that neither of the witnesses from the UKSA seemed to be aware of this. I doubt that it would be good value for the UK taxpayer but I would be happy to appear as a witness before the inquiry if the UK government would pay my travel expenses.
Eventually, people will realise tht there should be an owner-occupied housing component. Whether or not they accept the way it is done in the CPIH or the RPI, eventually something will have to be done about that.
House of Lords Economic Affairs Committee Call for Evidence
Today I received an email from the secretariat to the House of Lords Economic Affairs Committee with a Call for Evidence for their inquiry "The use of the retail price index (RPI)".
Some user group colleagues may have received a similar email.
The link to this call for evidence is at https://www.parliament.uk/business/committees/committees-a-z/lords-select/economic-affairs-committee/news-parliament-2017/rpi-inquiry-launch/
The pdf document included amongst the call for evidence documentation states –
"We are looking to hear from as diverse a range of views as possible-if you think someone you know would have an interest in contributing to the inquiry, please do pass this on to them."
User group colleagues may wish to submit evidence.
All the best.
Thank you for this. I hadn't received the e-mail you did but on your advice I had already written to the secretariat for the Committee. It seems they are quite serious about not accepting documents not prepared strictly for the inquiry, as I found out. I think the restrictions on length of submissions are quite strict considering the complexity of the issues involved and the small number of people who would likely bother to write something.
I just added a document to the library that is essentially a comment on the testimony on the RPI issue before the Lords Committee thus far: "The RPI Follows an Accounting Approach, Not a Blended Approach, to Owner-Occupied Housing Costs". I would like to thank Gemma Keane of ONS for valuable information about transaction costs in the RPI. I also made very minor changes to a previous paper I had written about the same subject two years ago, which is also in the library. The changes in no way attempt to remove evidence of the not very great evolution in my views since I wrote the original paper. My new paper much exceeded the limit for submissions for this inquiry even though it only deals with a small part of the topic. I really do think that submitters are not being given enough space to express their views.
Good luck with your own submission if you are writing one.
Just to let User Group colleagues know that the Economic Affairs Committee hears from Deputy Governor of the Bank of England, Dr Ben Broadbent this Tuesday 9 October 2018 at 3.35pm.
Colleagues may also be interested that on the investigation home page it says –
"The deadline for written submissions has now passed. Contact the Clerk if you still wish to make a submission. Oral evidence ongoing."
There have been about 40 written submissions to date and I will admit to having only read a small fraction of them.
Doubtless some colleagues will have seen the House of Lords RPI Committee evidence from the Bank's Deputy Governor yesterday.
I have to say I was surprised and indeed rather shocked at the Deputy Governor's statements about the Household Costs Index (HCI). He does not follow John Pullinger's opinion that the CPI (and CPIH) are "macroeconomic" indices as compared with HCI as a microeconomic household-based index. Indeed, he said that there can only be a single "best" measure of consumer price inflation.
I had the feeling that he suspects the HCI somehow to be a competitor of the CPI – and perhaps a threat to it. This is of course nonsense. In our paper "Towards a Household Inflation Index" (2015), Jill Leyland and I clearly presented the HII as a separate index from CPI and with a separate purpose – in fact a purpose in line with the origins of consumer price indices, focussed on inflation as experienced by households. The Deputy Governor said several times that he was unclear what the meaning of "households' experience" was. I imagine he has not read our paper!
He went on to give as an example the treatment of interest payments, referring, in the now familiar manner of dyed-in-the-wool national accounts-oriented economists, that interest received and interest paid are merely two sides of the same coin – one is a plus and the other a minus - and neither should be included in a CPI. He does not seem to appreciate that from the point of view of the household, interest – and especially mortgage interest - is a major outgoing. Interest does indeed have a price: the rate of interest, and a rise in that rate represents inflation for the home-owner. My experience of householders' attitudes to interest-in and interest-out is that if they get a windfall of cash their first priority is to pay off some of their mortgage rather than invest it in a bank deposit. But for the Deputy Governor all this is seen as nothing more than a matter of "semantics" and "language". It is actually far more serious than that.
The Deputy Governor also surprised me with his reference to the lack of an OOH measure in the HICP/CPI. He did not appear to know that the decision to measure OOH in the HICP was taken many years ago (in fact when I was still leading the development of the HICP) – including the decision on how to calculate OOH on the basis of net acquisitions, i.e. mainly new housing plus acquisitions from other sectors. Most if not all EU countries (including the UK) are now producing such an index, and it seems to be only a matter of time before it will be included in the HICP/CPI. Presumably that will then be a trigger for the Bank to move to CPIH instead of CPI.
Obviously, across the economy as a whole interest in must equal interest out. This is irrelevant to any sensible measure of consumer inflation, for two reasons:
Thank you very much for telling us about Ben Broadbent testifying before the House of Lords Committee.
One would expect a Deputy Governor of the Bank of England to be much better informed about inflation measures than he was. He spoke about Eurostat taking a long time to decide how they would incorporate owner-occupied housing in the HICP. This is completely false. From March 1997 forward, the only option that Eurostat has entertained has been an OOH price index based on the net acquisitions approach. It is strange that Mr. Broadbent would say this, as the UK has been part of all four Eurostat OOHPI pilots, starting in 2000. The big impediment to its introduction in the HICP has always been data related. By December 31, less than 12 weeks from when Mr. Broadbent testified, Eurostat must publish a report on incorporating OOHPI in the HICP. If it doesn't establish a timeline for adding OOHPI to the HICP, it will almost certainly be for data-related reasons. Lord Lamont, who as Chancellor of the Exchequer, introduced inflation targeting to the Bank of England, asked Mr. Broadbent if it was important that the inflation indicator used be consistent with international practice (as I recall) and was almost certainly thinking about this forthcoming Eurostat report. Mr. Broadbent said it wasn't, but gave no indication he was aware of the report.
Even as a dumbed down version of the existing RPI approach to OOH, Mr. Broadbent's description was inadequate. He implied that the only components included in the RPI were mortgage interest and depreciation. He rejected mortgage interest being in any kind of consumer price series on the absurd ground that the counterpart of mortgage interest paid was interest received by savers, but he would not, presumably, exclude haircuts from a consumer price series on the grounds that the counterpart of spending on them is income to the hair stylist. Mr. Broadbent much prefers imputed rent as a measure of OOH costs, but seemed to have no awareness that conceptually imputed rent covers off both mortgage interest and depreciation, those components he dislikes, as well as dwelling insurance, ground rent, owner repairs and DIY materials, the RPI components he chose to ignore. Imputed rent does not cover off RPI components for estate agents' fees and other real estate transaction costs, which are not part of consumption expenditure in a national accounting sense. Presumably, Mr. Broadbent does not support their inclusion in a consumer price series, but it is a pity that Lord Lamont didn't get him to explicitly say so. The Eurostat OOHPI is supposed to include all transaction costs, including stamp duty, which is omitted from the RPI.
When Mr. Broadbent was asked to confirm or deny that the OOH component in the Irish CPI was just an index of interest costs, he professed he didn't know. A mortgage interest CPI is in fact the only OOH component in the Irish CPI, based on current variable mortgage rates and a 20-year moving average of house prices, so it is quite similar to the mortgage interest component that Mr. Broadbent denounces in the RPI. Ireland is the only country with which the EU shares a land border, so you would think that Mr. Broadbent would have at least a basic idea of how the Irish CPI is measured. Possibly he does, and simply did not want to admit that the Irish CPI, unlike the UK CPI, directly incorporates changes in house prices.
Thank you again, Arthur.
I waited until the text of Dr Ben Broadbent's evidence was published before posting as I wasn't sure whether or not he had conflated two different versions of time reversal. I suspect that Dr Broadbent has done so but what he meant is not entirely clear –
The brief thoughts below are not intended to be rigorous –
The Fisher version implies that time can flow both forwards and backwards. This does not fit with our experience of time and is analogous to the sci-fi world of, say, Dr Who. It is also not consistent with the physics of the second law of thermodynamics.
There is no reason why an index formula should not be symmetric in time – most theories in physics are symmetric in time – but arguably it is unreasonable to condemn Carli because it follows the second law of thermodynamics and is not symmetric in time. Why should a decision on the choice of index formula be dependent on a phenomenon that will not be observed in the real world?
Also from an algebraic perspective the Fisher version involves the inversion of a price relative and because Carli is implemented in the RPI with a base month the denominator must always remain the base month price. Arguably inverting a price relative is not meaningful in the RPI context.
The second version does fit with our experience. Sets of prices can return to their earlier values and in that case the expectation is that the index should also return to its previous value. All three elementary aggregate indices – Carli, Jevons and Dutot – satisfy this version if only because they use a base month to calculate the indices.
In reality most price indices are more complex and, for example, quality change and chain linking affect outcomes. If there is a non-comparable replacement then a new base month price is imputed. This implies that if the same set of prices reoccurs after a non-comparable replacement then because the base month data has changed the index does not revert to the previous value – all the elementary aggregate formulae in the RPI and CPI are similarly affected. This can be interpreted as the prices are the same but the basket has changed. The process of chain linking has similar implications but it is far less likely that the larger number of prices involved would all return to the same values.
To conclude the criticism made by the deputy governor of the use of Carli within the RPI is unconvincing at best.
Link to text version of Dr Ben Broadbent's evidence
Thank you Shaun for your useful post today. It seems that the initial post which I made after the Broadbent session has become detached from the current HoL thread. I don't know if anybody can sew it back in?
I have added a document to the library on the same subject as yours, the inquiry into the use of the RPI. Thank you again for your help with it.
Yes – an excellent and well-focussed note by Hetan.
Thank you for pointing this out. The formula effect as measured by the difference between the RPI and the RPIJ has also gone down. It was -0.7% in December, but has been -0.6% for the last two months. Since the RPI inflation rate in February was 2.4%, unchanged from January, this means that the RPIJ inflation rate was 1.8%, unchanged from January. If the ONS has refined its methods in ways that permanently reduce the formula effect, it would be a welcome development, but the series has had its ups and downs in the past.
Even though the inflation rates for components like housing depreciation and dwelling insurance have come down a lot the RPIJ inflation rate still exceeds the CPIH inflation rate. In fact, to find a period when the RPIJ inflation rate was below the CPIH inflation rate for more than a couple of months one has to go back to the period August 2008 to December 2009, i.e. the time of the financial crisis.
In fact, the CPIH inflation rate is only as high as it is because of the decision, inexplicable on the face of it, to incorporate council tax in it. Council tax only changes once a year in April, so is still registering a 4.9% annual inflation rate, its highest since March 2005. The CPIH inflation rate excluding council tax would only be 1.7% in February, up from 1.6% in January and it would be really hard to sell an inflation rate so much lower than the alternatives as the headline measure.
On March 5, in his testimony before the House of Lords Economic Affairs Committee, Governor Carney would have us believe that he favoured an inflation measure that represented consumption pur et dur, and CPIH is that kind of measure. But it obviously is not: the scope of consumption expenditures in the National Accounts doesn't include council tax any more than it includes the real estate transaction costs whose presence in the RPI Governor Carney finds inappropriate. (He never questioned their presence in the Canadian CPI.)
The ONS should replace the CPIH with the RPIJ as the deflator of nominal wages. In the long term, a monthly Household Inflation Index would be the optimal deflator, but the ONS still seems to be a long way from turning the experimental HCIs into such an index.