Be careful of fiddling with the inflation figures
Disposable cameras are ejected from the government’s inflation calculations and cereal bars inserted: it may reflect our changing lives but it distorts the statistics. The 2010 retail prices index cannot properly be compared with the previous year’s if the components keep changing.
And if the stats aren’t truly comparable then they can’t be trusted. As the RPI is still used to upgrade state benefits and pensions and to calculate the return on indexed borrowing – gilts as well as mortgages - these figures matter.
The substitution of garlic bread for pitta bread or mineral water for fizzy drinks may be relatively small out of the 650 items in the indices, but the treatment of housing costs produces a bigger distortion.
The consumer prices index, the measure the Bank of England uses as its target to control inflation, excludes housing altogether, leaving transport the most important component with 16 per cent of our spending, followed by recreation, restaurants and, at 11 per cent, food.
But the RPI, which does include housing costs, is now to be adjusted to smooth the effects of interest rate changes by incorporating other mortgage products, including fixed rates. So today’s changes might not appear in the index for a year – and last year’s changes could come through only now.
In practical terms, it means that the long period of negative inflation the RPI recorded last year is now almost wiped away by the statistical revisionists. Prices fell only 0.5 per cent under the new methodology compared with the 1.6 per cent reported at the time. And the period of deflation has been retrospectively halved to four months.
Such changes undermine confidence in the statistical office as an independent provider of trusted data. It does not matter whether there is political manipulation in such changes; if people lose faith in them the damage is done.
March 17th, 2010 at 8:08 pm
The 2010 retail prices index CAN be compared with previous years. Price
indices always use a link-month, where both the old and new basket of goods
are collected. The index figure given by the new basket is then
chain-linked onto the the old basket, to give a continuous measure which
can easily be compare year on year. It’s all basic index number theory, which perhaps the author should have checked before posting. Also, the RPI has NOT been retrospectively revised, it is a complete
fallacy to say that is has. The new methods will only affect the calculations going forwards. Again, probably something the author should have checked.