Advisory panels on consumer prices

On 22 January 2016, the National Statistician’s Advisory Panels on Consumer Prices (APCP), Technical and Stakeholder, met separately to consider their advice to the National Statistician on his proposals to the board of the UK Statistics Authority for changes to consumer price statistics.

This note constitutes the combined advice of the two panels as discussed at these meetings and developed through subsequent correspondence. Panel members have had the opportunity to comment on drafts of this note. Where minority views have been conveyed by members these have been included¹. The panel chairs (Dame Kate Barker for the Stakeholder Panel and Nick Vaughan for the Technical Panel) have finalised the advice.

The panels considered a paper written by ONS that set out draft proposals and asked panel members to provide advice to the National Statistician on a number of key issues. This note follows the structure of that paper.

Notes for Context

1. Unless specified otherwise, the minority views in this note constitute the combined views of Jill Leyland, Andrew Sentance and Geoff Tily of the Stakeholder panel, plus John Astin of the Technical panel.

In providing this advice the panels acknowledged that the present UK landscape for consumer price statistics is complex, with a range of measures, two distinct legal frameworks and a number of areas of methodological contention to deal with. This is not necessarily the starting point one would choose – it is however the starting point we have.

The stakeholder panel noted that this is an opportunity to develop existing or new measures of price change that will meet a range of user needs and are better understood than at present. However, there was also acknowledgment from both panels that it would never be possible to produce sufficient measures to precisely meet every need. Furthermore, members of the technical panel were of the view that it was important for ONS to focus on one particular approach to measuring inflation and “get that right” – acknowledging that no single measure could precisely meet every need.

The panels noted that there should be a distinction between differences in inflation rates resulting from the use of different statistical methods, and the differences resulting from the different perspectives on what constitutes price change.

In developing their advice, panel members took note of the outcomes of the recent consultation on consumer price statistics and wider work including the Johnson review, the advice of the previous Consumer Prices Advisory Committee and ONS research. Doubts were expressed by some members about how widespread knowledge of the consultation had been among sections of the user community. Certain members felt the treatment of owner occupiers’ housing costs in CPIH had not been described in sufficient detail in the consultation to enable less expert respondents to reach an informed view on the topic.

A number of panel members expressed concern about the limited time the panels have been given to provide advice.

A minority of panel members felt particularly strongly about the impact of time constraints, and considered that this had not allowed the panels to give proper consideration to all the complex issues involved. The same members stressed that this was a “once in a lifetime” opportunity to set a new course for measuring consumer prices in the UK. In taking this opportunity, these members stressed the importance of not rushing to a decision and not ruling out options without due consideration.

A1 – Should there be a main measure of price change across the economy?

Both panels discussed what ‘main’ meant in the context of an inflation measure and whether this was a useful term. On balance, both panels expressed concern over the clarity and relevance of the concept of a main measure as defined in the draft proposals they considered. Members of the stakeholder panel in particular were of the view that there should be further debate about user needs as it was not clear whether a single measure could sufficiently capture the key uses and thus be reasonably designated as ‘main’. Specifically, a number of panel members questioned whether a single measure could adequately fulfil the differing needs for macroeconomic management and for indexation (for example, the uprating of benefits). In summary, the panels were of the view that the use of the term ‘main’ was not helpful to users and that ONS should reconsider its use.

Panel members were clear that even if the ONS were to designate a main measure, this should not drive out the perfectly reasonable requirement for a range of measures for different purposes – notably to improve understanding of households’ experiences. This was important to distinguish from ‘inflation shopping’, where users focus on the measure of price change which best suits their financial interests. Some members noted that the effectiveness of a measure (or measures) as ‘main’ would come about through acceptance and use of the measure rather than any designation by ONS. Furthermore, the panels felt that it would be important for ONS to set out in its publications what concepts the different inflation indices were capturing – so that users could judge which measure best met their needs.

Many panel members remained unclear on the meaning of ‘price change across the economy’, and there was a sense that ONS has more to do in explaining this concept or consider renaming it. There was discussion on whether this concept incorporated the ‘cost of living’. There was agreement that it did not cover the formal, economic, definition of the term (i.e. the cost of maintaining a constant standard of living (or utility) over time). However, there was acknowledgement that ‘cost of living’ is often used in different ways. A member noted that the panels should discuss and resolve what this term meant in practice and what measures could meet this definition.

A minority of panel members¹ believed there were clear user needs for two main measures, one designed to reflect price change as experienced by households (in their view, the original aim of consumer price indices including the RPI) and one designed to meet the needs of “macroeconomic management and international comparison including inflation targeting (such as the HICP/CPI)". One of these members felt the use of the phrase “price change across the economy” was dangerously confusing and should be replaced by a more conventional and clearer definition.

A2 - Why CPIH should be the main measure of price change across the economy

A substantive majority of technical panel members supported a move to CPIH using the rental equivalence approach to measure owner occupiers’ housing costs as the focus for ONS reporting on consumer price inflation. They noted that its construction was broadly in line with international best practice and current statistical thinking on the measurement of price change.

The stakeholder panel took a different view, noting that it was not clear that CPIH as currently designed would be suitable to meet a broad range of user needs – indeed, it was unclear if it were possible for any one measure to meet these needs.

Both panels highlighted three priorities for building confidence in CPIH. Firstly, members felt that the technical processing issues with the measurement of owner occupiers’ housing costs in CPIH that resulted in the loss of its National Statistics accreditation had to be fully and publically addressed. Secondly, CPIH had to regain its National Statistics accreditation. Thirdly, the measure needed time to ‘bed in’ and build traction. The panels stressed the importance of ONS communicating the appropriateness of CPIH (notably the OOH component) from a user perspective in order to achieve these priorities.

A minority of members did not view these three priorities as adequate conditions to enable CPIH to become the main measure of inflation published by ONS. They shared a view that the Johnson review’s recommendation that CPIH should become the main measure had been accepted “without proper debate and with little reasoning” either in the review or subsequently. These members also felt that there had not been adequate discussion of user needs (current or evolving) in the Johnson review nor (so far) by the panels. These members were of the view that CPIH is currently not in a position to be adopted as the main measure of price change across the economy. Whilst acknowledging that it should be possible to resolve the issues that caused CPIH to lose its National Statistics accreditation, these members noted a fundamental need to build public trust in the measure. These members set out a number of steps over and above those endorsed by the majority of panel members they believed necessary to ensure CPIH could be accepted as a reasonable estimate of price change as perceived by the public.

1. A complete review of the appropriate method for measuring owner occupiers’ housing costs – noting concerns over the ability of the rental equivalence method to gain public credibility, the reliance on an imputed method for such a large proportion of the index, and the different evolution of rents from house prices.

2. An examination into the choice of elementary aggregate formula used in CPIH and a resumption of the research into dealing with clothing prices and other goods which have the largest contribution to the formula effect. These members felt this work was needed to counteract a perception and (in the view of these members) a very possible reality that CPI and CPIH underestimate inflation in areas such as clothing and to improve the accuracy of all indices.

3. An active programme to build credibility in CPIH due to its de-designation as a National Statistic and current lack of scrutiny by the user community.

4. A clearer articulation of the purpose of a main measure of inflation. These members felt ‘price change across the economy’ needs to be clarified as it seems to imply more than just consumer prices.

5. A plan to develop an overall Household Inflation Index which could also act as a check on how close CPIH is to households’ experience of price change.

6. A review in due course of certain other areas of CPIH, such as the treatment of insurance, where these members believed current practice could be considered questionable.

These members felt that the first four of these steps should be completed before CPIH became a main measure. It was the view of these members that a main index will very probably be used for uprating purposes by some entities, whether in the public or private sectors, as well as for inflation targeting and other economic management. These members noted that this placed a long term focus on the index, meaning that the cumulative impact of any issue would be greater than it is for short term considerations.

Two of these members felt particularly strongly that the formula effect issues required further consideration before the optimal way forward should be agreed.

On the subject of traction, there was debate around whether a measure could be accepted as the focus for ONS reporting when it was not the Government’s inflation target measure and, conversely, whether it was desirable for the inflation target to be based on a measure which was not the ONS focus. Decisions on the choice of target are, of course, for the Government and are outside the remit of the UK Statistics Authority, ONS, and the panels. However, a majority of panel members argued that CPIH could be regarded as a potential option for the Government’s inflation target measure once confidence and communication of CPIH was established.

After some discussion, the panels reached a majority view that they supported the inclusion of council tax in CPIH. The panels noted that the arguments previously expressed for and against its inclusion were finely balanced. On the one hand, members noted that council tax reflected a large housing related cost to households and so should be included in the housing component of the index. Members also commented that its inclusion, while justified on other grounds, would also help CPIH gain credibility among the public. On the other hand, a technical panel member noted that its inclusion was (in their view) contrary to international best practice for a consumer price index because it was a tax rather than the cost of consuming a service. The same member noted that the inclusion of council tax would however, in their view, be consistent with international best practice for a cost of living index. Panel members were of the view that, should council tax be included, this should happen before CPIH becomes the focus for ONS and that the back series of CPIH should be revised to include council tax.

A technical panel member disagreed, judging that it still important to be consistent with international best practice, and that council tax could be more appropriately captured in an household groups' analysis.

The panels expressed concern over the name CPIH, with some members feeling this made it sound like a derivative measure (a parallel was drawn with RPIX and RPI). The panels considered whether CPIH should be renamed CPI in due course, but noted that this could be confusing and potentially presented legal complications (for example, where CPI is specified and used in contracts). The panels concluded that the naming of the indices was an important topic that merited further consideration and that users should be given plenty of forewarning of any change in name in order that any transitional problems could be signalled and resolved.

The governance of CPIH was briefly discussed. Panel members noted that respondents to the consultation were generally not in favour of legislation. A member of the stakeholder panel questioned whether this was the right approach, noting that users who required long-term consistency in a measure of price change would need sufficient reassurance that the Authority was committed to the continued and consistent production of CPIH. This would be particularly true if CPIH index-linked gilts were to be issued.

Notes for A – Measuring prices across the economy

1. This view was expressed by the minority group as named above, with the exception of Andrew Sentance who does not support this view.

There was general support from both panels for the value of producing measures of the impact and experience of price change for different household groups.

The panels noted that this type of measure would be useful from a public policy perspective, particularly when considered alongside comparable measures of income and expenditure. The panels felt that there should be a clear distinction between these measures and the CPIH or CPI, reflecting the different perspective on which they were based.

There was some discussion on the technical composition of these indices. The inclusion of interest payments (notably mortgage interest payments) was supported in the context of measures for different household groups subject to the inclusion of interest received (for example, from savings) in comparable measures of income. The potential inclusion of interest in the form of financial intermediation services indirectly measured (FISIM) was also raised. The subject of whether households should be equally weighted was discussed. The panels broadly supported the principle of an initial annual analytical publication of this type of measure but noted that the frequency of the release should be kept under review. The panels encouraged ONS to commence publication sooner rather than later.

The concept of an aggregate Household Inflation Index (HII) was discussed. There was agreement with the proposal for ONS to conduct research into the concepts raised in the ‘Towards a Household Inflation Index’ paper (Astin and Leyland, 2015). Most members agreed that the measures of the experience and impact of price change for different household groups was the right starting point for empirical testing and credibility building for the results of this research.

A minority of members felt that, while there was clear and agreed value in developing inflation measures for household groups, these were not a substitute for a single measure aimed at measuring inflation experienced from the household perspective, and capable for use for uprating and informing wage negotiations. These members supported the development of a HII alongside the use of CPI or CPIH, noting that this could reduce the use of RPI in the future.

C1 – Publication of RPI data

The panels broadly supported the proposal for limiting the future publication of RPI data to the aggregate RPI, its components, and RPIX. There was some consideration of continuing publication of RPIJ to help users understand the scale of the ‘formula effect’ in the RPI. However, there was agreement among a majority of members that publication of RPIJ was not essential should ONS improve the reconciliation of the differences between RPI and the main measure of price change across the economy. ONS noted that they were committed to doing this.

There was a minority view that RPIJ should be maintained as it is the “least worse” of current household-based indices and held National Statistics status. These members noted that RPIJ held much in common with the proposed Household Inflation Index and could be a ‘staging post’ in developing the latter index¹. These members also noted that RPIJ was an analytical tool but does not resolve the formula effect issue.

C2 – Principles for maintenance of the RPI

Both panels supported the clarification of the definition of ‘routine maintenance’ for the RPI as all changes required to continue production of a consistent, fit for purpose RPI. The annual update of the basket and weights, computer systems upgrades and improvements to data validation and quality assurance methods are all considered to be routine maintenance.

The discussions on the principles for developing the RPI focussed on the potential for future improvements to alter the size of the formula effect. Panel members considered whether ONS should adopt a clear stance to maintain or reduce the size of the formula effect. Panel members did not reach a conclusion one way or another on this point. However, there was general agreement that ONS should conduct thorough assessments, including consultation, about the likely impact of any change to the formula effect before proposing it for implementation.

In summary, there was broad support for the principle for the development of the RPI as set out in the draft proposals paper. Namely that, with due consideration to the requirements of the Statistics and Registration Services Act 2007, ONS will only incorporate improvements to the RPI if to not do so would inhibit the improvement of the CPI or CPIH. However, the support was caveated on ONS conducting thorough impact assessments of any proposed change on the path of each index and the size of the formula effect gap.

A minority of members noted that the RPI continues to be widely used in contracts and indexation, is vital for index linked bonds and is the longest index available for measuring consumer prices. These members were of the view that RPI would continue to be used for many years to come and should therefore be as accurate as possible, given the constraints on changing it. These members viewed a policy of ‘managed decline’ of the RPI as irresponsible and noted the importance of resuming work on the size of the formula effect. These members expressed strong reservations about setting aside a measure of inflation that still enjoys public recognition and confidence until there is another proven and accepted measure that can meet its original aims. Finally, these members noted that they viewed the ONS as having an advisory role to play in commenting on the practice of ‘index shopping’ by government.

Notes for C-The RPI

1. Jill Leyland abstained from commenting on RPIJ having declared a conflict of interest.