What’s that got to do with the price of fish? The growing gap between inflation and the perception of inflation

Last week’s news that the Consumer Prices Index (CPI) rate of UK inflation has dipped below zero for the first time since official records began in 1996, and for the first time since 1960 based on equivalent historical estimates, caused a range of reactions amongst commentators and politicians.

The Chancellor, George Osborne, was keen to emphasise that this should be termed “negative inflation”, a brief spell of decreasing prices, rather than the more damaging “deflation”, implying a longer term period of falling prices and economic stagnation such as what Japan went through during its “lost decade” in the 90s. Other commentators pointed to the fact that falling prices should give hard pressed households a break.

Households, however, presently have a very different view of inflation than the view of Government economists. The Bank of England carries out thousands of interviews with members of the public to gauge their attitudes towards inflation and other economic variables. They now make public the raw data of public responses to the questionnaire, making it possible to gauge what the public perceives inflation to be, and this can be compared with the official published inflation rates.

Perhaps unsurprisingly, since the survey began in 2003, the public have always perceived inflation to be higher than the official CPI rate. Bar for a period of negative inflation associated with lower housing costs in 2009, from 2003 to 2011 public perception of inflation was in line with the older inflation index, the Retail Prices Index (RPI).

Inflation and Perception

Historically, the RPI CPI measure of inflation has been lower than the RPI CPI. Part of the reason for this is that the RPI includes housing costs such as owner-occupiers’ depreciation, council tax and rates, and mortgage interest payments (see the Office for Budget Responsibility’s paper if you are interested). However, the most significant driver of the difference is mathematical; the RPI measure uses the familiar arithmetic mean method for calculating averages (sum all the prices in a category, then divide by the number of prices n). The CPI uses the geometrical mean method to calculate average prices; multiply all the prices together, and then take the nth root.

To take a practical example, the following are actual prices used in calculating CPI for December 2014, namely the prices of a pint of lager in independent bars in Northern Ireland:

£3.80      £2.15     £2.59     £2.80     £3.20     £3.30     £3.40     £3.45     £3.50     £3.60

Arithmetic mean: (£3.80 + £2.15 + £2.59 + £2.80 + £3.20 + £3.30 + £3.40 + £3.45 + £3.50 + £3.60) / 10 = £3.18

Geometric mean: (£3.80 x £2.15 x £2.59 x £2.80 x £3.20 x £3.30 x £3.40 x £3.45 x £3.50 x £3.60) ^ (1/10) = £3.14

What seems like a dry (and frankly dull) mathematical distinction actually can actually make a huge difference to people’s livelihoods. Pensions, for example, were changed from being linked to RPI to be linked to CPI in 2010. According to consultancy firm Aon Hewitt, this change of index could actually lead to the overall value of pensions being reduced by 25 percent. The uprating of benefits has also been impacted by the change of inflation index from RPI to CPI.

One might well conclude that a principal objective of this change was to save the Government money, especially considering that the new student loans brought in by the coalition Government will be linked to RPI, meaning repayments will be higher than if they were linked to CPI.

So, what is causing the gap between perceived inflation and actual inflation over the last few years? The Inflation Attitudes survey data is not the only interesting public domain dataset available to data and finance nerds. The Office for National Statistics now publish the millions of individual products and prices that go into producing the CPI index. If, for example, you have a burning desire to track regional variations in the price of driving lessons, or yoghurt, or “small caged mammals”, or any of the other hundreds of items used to derive the headline CPI rate, then go right ahead.

Recalculating the headline CPI rate from the base data is difficult, so I decided to roll my own alternative inflation measure from the price data to see if I could solve the puzzle of why people believe inflation is so much higher than the Government says it is. Prices are segmented by product, UK region, and the type of shop (i.e. whether it is a chain or an independent retailer). I took the geometric mean of all the prices in each bracket, and then compared with the geometric mean of the prices in the same bracket in the previous year. I then used the regional weights and product weights to weight the results. Essentially, it should be a measure of how much the prices in the CPI basket have changed since the same month in the previous year.

Various Inflation Measures v2

All this shows, essentially, is that measuring inflation is hefty amounts of art and subjectivity to a dollop of science and objectivity, and public perception of inflation is often far removed from those of Government statisticians. The crude, amateurish approach that I used to calculate inflation is frequently closer to public perceptions than the official CPI index.

The “basket” of goods used to calculate CPI is also, by necessity, an approximation of the “typical” consumer’s consumption choices, and can never reflect individual circumstances. Take, for example, the weighting given to nappies in supermarkets, which was given a weighting of 0.64, and for which prices increased by 35% in 2014 in Northern Ireland. On the other hand, take the weighting given to the price of pint of lager in a chain pub, which was given a weighting of 5.46, and which (apparently) reduced by 9% over the year.

The change in beer prices will therefore have more than eight times the impact than the change in price of nappies. Although I remember fondly the days when the price of a pint had more of an impact on my own finances than a box of nappies, this hasn’t been the case for some time, and my own personal inflation rate will be higher than the “generic” CPI as a result.

There are many other reasons when perceived inflation might be higher than actual inflation. One reason could be that we tend to gauge inflation by fluctuations in inflation by those goods we buy the most, such as petrol or pints of milk, and if these fluctuate over the year we tend to remember the price increases and forget the price decreases. Another is the fact that, since 2011, pay rises have not kept up with inflation, so those answering the questionnaire may simply think prices are increasing due to their decreased spending power.

When it comes to the price of fish, or indeed the price of anything, the situation is complex and the relief afforded by a (presumably temporary) fall in the CPI rate will not be felt by everybody. The gap between what is felt by consumers and what is calculated by Government statisticians grows ever wider.

, ,

  • Korhomme

    Fascinating info, SoD, thanks!

    And I’m with you, the price of nappies no longer concerns me, but the price of a pint does. And isn’t this the problem with these indices? They relate to a ‘basket’ of goods that the ‘average’ consumer buys, and this is an entirely artificial measure. As an average, it is a ‘construct’, a theoretical model; useful for trends across the measures, but of so little relevance to the DINKYS or the silver surfers. It’s no surprise that individual’s experiences diverge so much from the mean; but then, none of us consider us ‘average’, do we?

  • AndyB

    I think you mixed up CPI and RPI in “Historically, the RPI measure of inflation has been lower than the CPI” 🙂

    A key driver in perceived inflation is also that supermarket prices are not falling. Offers come and go, and milk and bread are regular loss leaders, but it’s a rare household that will tell you their supermarket bill is as little as a couple of quid more than the same time last year.

  • hugh mccloy

    nice work, maybe for a further write up, negative interest rate loans, been catching up on some of the finer details of new articles

  • T.E.Lawrence

    “Pensions for example were changed from being linked to RPI to be linked to CPI in 2010” Is this both State and Private Pensions ? I thought Private Pensions was still linked to RPI ?
    Agree with your statement “One might well conclude that a principal objective of this change was to save the Government money ! Good article Salmon.

  • Sergiogiorgio

    As important as the price of a pint is, surely the key driver is oil price, as it not only “should” correlate to the price at the pump but also the costs of moving all the goods that we buy. Because, I would contend, a racket is being played between petrol and oil price, people do feel inflation, at best, remains static.

  • AndyB

    Only the “Guaranteed Minimum Pension” element of private final salary related pensions has to be index-linked by law, and much of that index-linking is paid with a person’s State Pension. Otherwise, I believe private pension providers, including annuity providers, are pretty much at liberty to increase pension payments or not as they please, unless the rules of the individual scheme specify otherwise. A quick google search reveals that a lot of pensioners are on fixed-rate annuities because the regular income would be reduced by such a degree to pay for index-linking.
    Public Sector Pension schemes were indeed changed from RPI to CPI to save money.

  • T.E.Lawrence

    Thanks AndyB for that information.

  • Megatron_

    Not quite that simple I am afraid….

    Guaranteed Minimum Pension ceased to accrue after 1997 so it is generally a very small amount so I will ignore it.

    For private sector defined benefit schemes there are two places where indexation can directly affect pensions – increases to benefits for members who left before retirement in the period up to their retirement and post retirement increases.

    For members who leave their whole benefit is subject to indexation in the period up to retirement. The basis for this indexation changed from RPI to CPI in 2011.

    Similarly for defined benefit schemes post retirement increases switched from RPI to CPI in 2011. This generally affects all pensions accrued after 1997 in defined benefit schemes where were subject to indexation subject to a cap of 5% or 2.5% (depending on scheme).

    As an added wrinkle some of the above could be overwritten by individual schemes trust documentation if it specifically states RPI as the method which is to be used.

  • Megatron_

    I should have added that the point about annuity providers is correct in that they basically convert a pot of money into pension so they will give you whatever you want and change the price accordingly. The RPI -> CPI change was about saving the state and employers money.

  • T.E.Lawrence

    Thanks Megatron, as I thought a mess of confusion which I am sure does not suit the poor punter !

  • salmonofdata

    Oops! Thanks, I’ve fixed that now.