Phillip Howard — Written evidence (RPI0023)
1. Is the current situation regarding the retail price index is untenable?
1.1 Yes. I believe the production of two conflicting indices is counterproductive for the UK. It causes confusion, which in turn makes people less likely to devote time to the understanding of their lifetime savings. When organisations (i.e. trustee bodies) require understanding of inflation to perform their roles unnecessary time and money is spent on the topic. Bringing the UK in line with other countries with one main inflation measure would improve productivity; the great minds discussing the issue at the House of Lords select committee and many others will able to commit time to other important issues.
1.2 ONS should devote its time to providing the best official CPI index and not waste time on an index it believes to be inappropriate and outdated.
2. Should the retail price index be abolished?
2.1 No. This should still be published without the need for the complex construction by the ONS until the last RPI index-linked gilt has matured in 2068. This equates to roughly a further 600 monthly RPI prints! The work required to calculate RPI for the ONS (or other public body) would be to produce an Index according to a method similar to that shown below, deriving solely from the CPI. There would be no need to not spend time on collecting prices, considering basket constituents and regularly reviewing (explaining deficiencies of) calculation methodologies. These actions are currently required to support the production of the RPI.
3. If not, how should the retail price index be changed?
3.1 The RPI should be constructed by simply scaling up the change in CPI by fairly adjusting for the expected formula effect (“wedge”) as at 2018. An example methodology is presented below:
3.2 The consensus for the wedge means on average market participants (and trustees alike) expect that the year-on-year RPI change will be 1% higher than the year-on-year CPI change.
4. What would the implications be of changing or abolishing the retail price index?
4.1 A replacement of the nature presented in the previous question would mean:
(i) A simpler framework to judge UK inflation by;
(ii) less confusion for market participants and individuals around the measure of inflation;
(iii) enable an orderly move to CPI over time (as the calculation methodology of an index-linked Gilt’s Real (CPI) yield would become very straight forward given its current price)
4.2 Whilst the best time to provide one central UK inflation measure may have been about 20 years ago, the second best time is now!
5. Personal background
5.1 I have provided some personal details and thoughts to provide context to my responses to the questions posed.
5.2 Over the past 10 years I have worked with defined benefit pension schemes to advise on, and latterly to implement, the structure of liability hedging arrangements. In the main, these arrangements involve purchasing index-linked gilts (directly or via the use of leverage) to match a proportion of CPI or RPI linked liabilities.
5.3 In late 2012 / early 2013, I co-ordinated my previous employer’s response to the Consumer Prices Advisory Committee consultation. At the time it struck me that the Committee placed greater weight on the volume of individual responses than on the crucial nature of the methodology used to calculate the RPI and CPI, which was considered a secondary issue to the size of the formula effect (“wedge”). In my view, to generalise the outcome of that consultation, with the unintended Carli consequences maintained led to:
(i) a lifetime of improved incomes of individuals with RPI linked pensions; at the expense of:
- the UK government whose index-linked gilt coupon and principal inflation uplift were now expected to be higher for stock in circulation at the beginning of 2013 (n.b. future issuance have the wedge “priced in”)
- the funding of corporate pension schemes which provided RPI linked benefits.
5.4 The wealth transfer would have been even more pronounced if the UK government had not already transferred public sector pensions from an RPI uplift to a CPI uplift two years earlier (citing a negative year-on-year RPI change as the rationale).
5.5 The fact that corporate defined benefit pensions are now uplifted with a mix of RPI and CPI also causes issues for schemes looking to hedge their liabilities accurately or secure a bulk-annuity with an insurance provider. The lack of secure bonds with CPI uplifts (and therefore lack of market liquidity in CPI linked assets) means predominantly CPI linked liabilities are hedged with RPI uplifted index-linked gilts (or swaps) in an approximate manner. This is a reasonable strategy given a reasonably close relationship between the two indices although notably the relationship did not hold during the last financial crisis. When pension schemes are looking to secure their liabilities with insurers, the insurer penalises schemes with CPI liabilities as the insurer cannot very closely match the scheme’s liabilities with the by investing in CPI linked assets (they, the insurers, compensate for this by increasing their premium). This unfairly higher cost is due each insurer’s capital rules. This leave trustees with the often unpalatable option of (i) paying the higher insurance premium or (ii) in some notable instances going to the vast expense of offering members (including paying for financial advice across the membership) to exchange CPI benefits for fixed amounts ahead of a potential bulk-annuity purchase.