RPI is dead? Long live RPI!
Today, the Office for National Statistics (ONS) announced its decision not to proceed with its recent proposal to recommend an “improvement” in the calculation of the Retail Prices Index (RPI).
The RPI uses a number of formulae, one of which is called the Carli formula and this calculates an average (mean) arithmetically. The ONS proposed changing from Carli to a formula that calculates geometrically: this isn’t an improvement, it’s a change and the two aren’t synonymous.
The British Government, institutions and others are aware of the basis of the RPI and have priced their contracts accordingly. People have bought RPI-linked lifetime pension annuities based on the RPI being calculated with its existing formulae; to change those RPI formulae so that future increases are less than they would have been is effectively to defraud those people from receiving what they paid for.
The ONS has decided not to recommend a change in the RPI formulae; rather, the RPI will continue in its current form and a new monthly index, called RPI-J, also will be published. This enables existing contracts based on the RPI to continue as they were agreed, whilst new contracts can be based on whichever index the contracting parties agree to use.
The ONS are to be congratulated for this sensible decision.
For those who may be interested in such things:
An example of the difference between arithmetic and geometric averages:
The arithmetic mean of 24 and 6 is (24 + 6)/2 = 30/2 = 15.
The geometric mean of 24 and 6 is √ (24*6) = √144 = 12.
RPI-J alludes to Jevons, a geometric formula.