RPI-CPI: Yeah but am I bovvered though?

Tweet
I'd almost forgotten about the investigation by the ONS (Office of National Statistics) into the causes of the gap between the Retail Prices Index (RPI) and the Consumer Prices Index (CPI), until I was reminded about it this morning.

Now you may be wondering why mums in Surbiton are asking me about inflation indices on my morning school run. I am relieved to report that whilst Surbiton and its surrounds is famed for the high density of LDI specialists who call this part of South East England home, we are not yet bored enough to preach the LDI gospel to school-run mums.  Instead my memory was jogged by a question from a former colleague.

1.     Mind the gap

Here in the UK we have two indices which provide measures of UK inflation, RPI and CPI. The “RPI-CPI gap” refers to the difference between these two numbers.

2.     Not all gaps are created equal

CPI was introduced in 1996 and since 1997, when the first 12-month figure for CPI became available, the UK’s RPI-CPI gap has averaged 0.9 percentage points, i.e. annual RPI has been on average 0.9% higher than CPI.

To place this into context, when looking across the developed economies of the EU, the US and Canada, then the UK stands out as the only country to have such a large gap between its two inflation measures. Taking the EU, for example, countries in the EU produce both a HICP (Harmonised Index of Consumer Prices) as well as a national measure of inflation but, with the exception of Slovenia, none of these countries exhibit a difference between these two measures of inflation which is as large as the difference which prevails in the UK. It turns out that the UK and Slovenia both adopt two different calculation methodologies which creates this gap and which we explain in more detail below.

3.     “Am I bovvered though?”

Well I guess it depends on who you are. If you’re Lauren, then clearly not.

For pension scheme members, then if your pension is linked to CPI and CPI is systematically lower than RPI then pension increases are lower than they would have been had inflation increases been linked to RPI. This is relevant because some UK pension schemes have been able to switch to indexing pensions to be in line with CPI rather than RPI.

From a sponsor’s viewpoint, since CPI has historically been lower than RPI, the presumed impact is that being able to switch pension increases away from RPI to CPI should mean a reduction in the liabilities of the scheme. As we will see later on, the current review underway by the ONS may mean that the historical norm of CPI being below RPI, and sometimes significantly so (in excess of 1%), may not necessarily be the norm in the future so that we should not be too hasty in presuming reduced liability values.

4.     Why is the gap under scrutiny now?
CPI has recently become more important with pension, benefits and tax credits now being indexed in line with CPI. In addition, taxes and duties may also be indexed to CPI. Differences between RPI and CPI are therefore receiving greater scrutiny than ever before.

The increased RPI-CPI gap which emerged in January 2010 was the catalyst for the ONS to investigate the cause of the gap and the reasons behind its increase. In particular, since the importance of CPI is increasing, emphasis is being placed on the calculation methods for CPI.

5.     What aspects of the methodology are being looked at?

Specifically, the ONS identified that the main reason for the gap was the "formula effect".

Put simply, the "formula effect" arises because CPI uses a geometric average of prices whilst RPI uses an arithmetic average of prices ⁽¹⁾. A mathematical feature of these two averaging methods is that the geometric average is always less than (or equal to) the arithmetic average. This gap is wider, the wider the dispersion in the underlying data being averaged.

The average RPI-CPI gap since 1997 has been 0.9 percentage points, of which 0.5 percentage points can be ascribed to the “formula effect”.

Further, the main reason for the increase in the RPI-CPI gap from 2010 onwards could be ascribed to the “formula effect” and specifically as it applied to measuring clothing price inflation - the wider dispersion in clothing price inflation data was the primary driver for the increased gap from 2010 onwards. The wider dispersion meant that the CPI number, being a geometric average, was pulled down whilst the RPI number being an arithmetic average was not affected in the same way. The impact can be seen in Chart1 below.  

6.     But why has this “formula effect” only now become an issue?
In January 2010, changes to the measurement of clothing prices were introduced and it was these changes which resulted in an increase in the dispersion of the clothing price data being used in the RPI and CPI calculations. This then translated into a larger RPI-CPI gap as can be seen in Figure 2 below.

 7.     So why not just change the averaging methods to be consistent between RPI and CPI?
Regulations mean that this is not straightforward.  The CPI calculation is governed by European Regulations⁽²⁾ which are legally binding and mean that the UK can not make unilateral changes to the CPI calculation. For example, EU Regulations prohibit the use of an arithmetic average (of price relatives) for the purposes of CPI calculations – this regulation means that the CPI calculation could not be unilaterally changed to be consistent with the RPI calculation which uses such an arithmetic averaging method. The UK could introduce a separate national CPI index though.

On the other hand governance of the RPI comes from the Statistics and Registration Services Act. Changes to the RPI calculation must be carried out in consultation with the Bank of England who must opine on whether the proposed changes would be materially detrimental to the interests of the holders of index-linked gilts. If this is considered to be the case then changes will require the consent of the Chancellor of the Exchequer.

8.     So, what is being done by the ONS?

In essence the ONS is looking at ways in which they can reduce the dispersion in the data they collect on clothing prices. This would then, in turn, reduce the impact of the “formula effect” and hence reduce the RPI-CPI gap. The ONS is also researching which averaging method (e.g. geometric average, arithmetic average or some other average) should be used for averaging clothing prices ⁽³⁾. However, as noted above, changes in the averaging methodology are complicated by regulations governing the calculations of CPI and RPI.

And so, what’s the bottom line.
Don't delay hedging your inflation risk. The likely outcome of the ONS investigation into the RPI-CPI gap is a reduction in the gap from 2013 onwards. This will make it more attractive to hedge liabilities linked to CPI using RPI hedging instruments. This has (at least) two implications:

1) greater demand for RPI hedging instruments and (all else equal) a commensurate increase in their price; and

2) smaller demand for CPI-linked hedging instruments and hence lower likelihood of these ever being available at a sufficiently attractive price, so if you are waiting to hedge CPI liabilities with CPI-linked bonds or other instruments you may be, in the words of Bob Marley, waiting in vain.

************************************************************************
And now for something completely different and on the lighter side here's something from a book I am currently reading -“Damn Good Advice” by George Lois.

“When asked to create a poster by Gary Kasparov’s handlers for the World Chess Championship between Kasparov and his brilliant challenger Anatoly Karpov at the Hudson Theatre in Manhattan, I created the ultimate confrontation in the fierce combat of chess. Kasparov’s business managers insisted that Kasparov, arguably the greatest chess player in history, would be oblivious to it – and forbid me to show it to him. But after an argument lost in translation, I defied them, and when I presented the poster to the Russian chess genius, the white chess piece between his profile and Karpov’s hit him like an emotional illumination, and he gasped the words “Na Zdorovye, tovarich! Kasparov and Karpov nose to nose, and betveen them ah vite kveen!”

Great work must be presented to the person that has the power to accept your creations. The problem is that the underlings in any business or enterprise can always say “No” (and many times do) but have no power to say “Yes” – so you must get past them and present to the decision-maker!”

Notes

(1)  The CPI predominantly uses the geometric mean (also known as the Jevons index, named after the English economist who proposed this measure of averaging), supplemented in most cases by a particular form of the arithmetic mean known as the ratio of average prices (the Dutot index). The RPI predominantly uses a combination of the Dutot and the arithmetic average of price relatives (the Carli index). This leads to a difference between the two indices known as the ‘formula effect’. (Source: ONS)

Carli
Developed in 1764 by Carli, an Italian economist, this formula is the arithmetic average of the price relative between a period t and a base period 0.

Dutot
In 1738 French economist Dutot proposed using an index calculated by dividing the average price in period t by the average price in period 0.

Jevons
In 1863, English economist Jevons proposed taking the geometric average of the price relative of period t and base period 0. When used as an elementary aggregate, the Jevons index is considered a constant elasticity of substitution index since it allows for product substitution between time periods.

(2)  The HICP is specified in a series of legally binding European regulations. Regulation number 2494/95 of 23 October 1995 provides the legal basis of the HICP and a series of subsequent regulations signed between 1996 and 2007 define its construction and coverage.

The HICP regulations are drafted by the European Commission in conjunction with experts from Member States and the European Central Bank through an HICP Working Party. The work is overseen and approved by the Statistical Programme Committee made up of heads of EU National Statistical Institutes (ONS in the UK) and the head of the European Statistical Office (Eurostat). Like other Member States, the UK can influence the legislation but can not dictate it. Regulations normally require a qualified majority vote before they are brought forward for legislation.

Once introduced, the regulations are obligatory in every Member State. Eurostat undertakes an ongoing monitoring programme to evaluate compliance with the regulations. In December 2003, the National Statistician decided that the name of the UK version of the HICP would change to the Consumer Prices Index (CPI) in all National Statistics publications though there was no intention to develop the CPI differently from the HICP. (Source: ONS)

(3) Based on certain assumptions regarding consumer preferences, economists believe that if a price index is being used for cost of living adjustments, the averaging method to be used should reflect that when individual prices change within a year, consumers may vary their purchasing behaviour accordingly in order to maximise their satisfaction, or utility. These changes in purchasing behaviour can be measured as the constant elasticity of substitution. When the constant elasticity of substitution is close to one, this effect can be shown to be accounted for by the geometric average (Jevons), whereas the Dutot and arithmetic average (Carli) imply that quantities purchased remain fixed, regardless of any changes in relative prices. (Source: ONS)

Tweet