Every year the furore around rail fare rises reminds us how the government games measures of inflation so that it charges RPI but pays CPI.
But the noise over rail fares should not distract from the ongoing battle on inflation measurement being played out between the interests of employers and workers.
This is a battle between the Consumer Prices Index (CPI) figure targeted by the Bank of England and the Retail Prices Index (RPI), which remains the dominant benchmark for pay bargaining.
A percentage point difference between the two measures may seem like small beer, but it matters.
For the average full-time worker, a pay rise in line with CPI rather than RPI strips approximately £350 from their annual income.
And the government recently estimated that any shift from RPI to CPI indexing of pensions costs the average pension member £12,000.
So the fact that big guns such as Bank of England governor Mark Carney have called for the phasing out of RPI based on false charges levelled against it by the Office of National Statistics (ONS) is worrying.
The RPI was virtually undisputed as the benchmark measure of inflation in the UK for around 50 years after the Second World War.
CPI began to gain greater prominence from 2003, but it wasn’t until the early part of this decade that the RPI started to face attacks on its validity, when the ONS was pushed by Eurostat and the Bank of England to change the way in which it collected price data on clothing.
Ever since its inception, CPI has massively understated inflation for clothing, without appearing to cause any great concern to the ONS (or the Treasury and Bank of England for that matter).
But following their rushed price collection changes, most of the error in clothing shifted to RPI. Rather than responding by researching clothing price collection properly, the ONS then started a campaign against RPI which it has continued ever since.
This is despite the ONS facing overwhelming opposition whenever it has held a consultation or engaged in a public meeting on the topic.
We would of course have to accept the negative consequences of CPI for workers if the charges laid against RPI were indisputable.
However, the ONS has never convincing proved its claims, and its shifting justifications for CPI reflect shaky foundations.
First of all, we were told that CPI accounted better for “consumer substitution”, ie consumers shifting to cheaper alternatives when prices rise.
Then evidence emerged that led the ONS to accept that consumer substitution offers no basis for arguing that CPI is a more accurate measure.
Secondly, the ONS made much of the idea that a statistical tendency called “price bounce” meant the RPI was significantly exaggerated. Evidence then emerged that any such effect in the RPI is miniscule.
The justification now appears to be based on the argument that CPI is international best practice.
We believe this is a conflation of “best practice” with “common practice”.
The reason why CPI has emerged as common practice is partly because countries such as Australia and the USA converted to CPI on the now discredited consumer substitution argument.
Defending RPI as the most accurate indicator of the cost of living changes facing workers is on much more solid ground.
It has always been apparent that RPI is based more tightly around the spending patterns of workers than CPI, since it excludes most expenditure by pensioners dependent on state pensions, tourists and the ultra wealthy.
And it is well known that RPI includes housing costs, whereas CPI excludes this most fundamental of household expenditures.
The ONS have now developed CPIH to incorporate housing, but the approach is controversial.
However, most of the difference between RPI and CPI is not down to whose costs or what costs are covered but to the statistical technique used to calculate them.
These statistical techniques are analysed exhaustively in Dr Mark Courtney’s report , which provides the theoretical bedrock for the defence of RPI.
The key finding of Courtney’s work is that almost 80% of the difference between RPI and CPI is caused by the fact that CPI under-estimates the real change in the cost of living facing workers.
Work by Courtney and others has shown that the problems in measuring clothing inflation could have been resolved by adjustments to the methodology for that specific range of goods.
But instead, the ONS has used issues it created by changing the price collection techniques for clothing (which accounts for just 4% of the total value of RPI) as a basis for attacking RPI as a whole.
UNISON acknowledges that other measures of inflation are needed for other purposes, most notably a macroeconomic indicator covering all forms of household.
However, as a measure of changes in the cost of living facing workers, RPI remains the most accurate indicator.
It seems unlikely that the ONS will do a volte face any time soon and restore the “National Statistic” status that it stripped from RPI in 2013.
However, the stronger the trade union defence of RPI, the harder it will be for the ONS to deliver either a quick death to RPI through ceasing publication, or a slow death through neglecting maintenance of RPI and thereby allowing confidence in the measure to wither.
If we allow RPI to meet such a fate, we will have permitted the most gigantic act of statistical pickpocketry – delivering gifts to employers at the expense of workers.
Want to hear about our latest news and blogs?
Sign up now to get it straight to your inbox