The official statistics body has today launched a consultation exercise to find out what all the many parties interested in the calculation of inflation rates think about the UK’s current two-rate system.
The two rates in question, you will recall, are the older Retail Prices Index (which includes housing costs) and the newer Consumer Prices Index, which does not. For most of the last decade, RPI has generally been about 0.5% - 1% higher than CPI (although occasionally, confusingly, it is the other way round). The ONS wants re-jig the way RPI is calculated to make it follow CPI more closely. Simples.
Or not - for such an apparently arcane subject, this consultation promises to be a pretty lively exercise. For starters the argument between stats geeks over the mathematical minutiae of the formulae used to compile the indices verges on the quasi-religious. For main course, there’s the small matter of billions in RPI-linked government bonds held by pension funds and savers, which stand to lose out from the proposed changes. And for pudding, the £200bn plus benefits budget, which wee Georgie is so keen to trim. Rises in many benefits are RPI-linked so any tweak which brings it closer to CPI could save the government billions, at least in terms of the future costs of rises yet to fall due.
The ONS fortunately has a good fig leaf that enables it to ostensibly keep out of all these murky politics - standardisation. RPI is an insular old UK-only measure dating back to WWI whose time is past. CPI on the other hand is a modern international standard measure which surely all good citizens of the global economy should prefer?
As for the reasons why the two indices differ so much, well this is where it get technical. RPI uses arithmetical price averaging - add up the total and divide by the number of separate items. CPI uses rather more sophisticated geometric averaging (multiply the prices of n number of items and take the nth root as the average). The resulting divergence is fairly stable and called the Formula Effect.
But there’s another bigger issue with the maths used for RPI, say its critics. The Carli Formula to be precise, yet another form of arithmetical averaging (you never knew there were so many ways to do it did you?) applied to some of the goods in the RPI basket. It results in an inherent upward bias, so if prices rise in the course of a year and then fall back again, the RPI stays up.
CPI apparently avoids this problem by using the Jevons or Dutot averaging formulae. (No, we’d never heard of them before either). That sounds like a win for CPI, except that there are other stats bods who argue that these alternative methods swap positive for negative bias, under- rather than over-estimating the rise in inflation. So it seems that all the available mathematical models aren’t quite perfect, who’d have thought it?
No prizes for guessing what the results of the consultation will be, private investors and pensions funds can bellyache all they like but the prospect of a big win on the benefits budget will be too much for the cash-strapped Government to resist. Expect near rate-parity to be introduced in due course.