-Adds Detail To Version Transmitted At 0856 GMT

LONDON (MNI) – A report from an international expert, commissioned
by the Office for National Statistics, has recommended axing a formula
used in the retail price index in a move which could substantially lower
the RPI.

While the proposed changes are technical in nature, lowering the
RPI has a potentially substantial impact on index-linked gilts and the
cost of UK debt payments. The report issued Monday, from Walter Erwin
Diewert, comes out firmly in favour of scrapping the Carli formula.

The summary of Diewert’s report says its “most important
recommendation for improvement is … the Retail Price Index (RPI)
should drop its use of the Carli index as an elementary index and
replace it by either the Jevons or the Carruthers Sellwood Ward and
Dalen elementary index.”

The Carli formula is an average which has led to RPI clothes
inflation being markedly above its equivalent in the CPI.

Alan Clarke, economics director at Scotia, has noted that scrapping
Carli could drag RPI down 0.3% year-on-year forever, although obviously
the effect is dependent on the choice of its replacement.

Diewert, who made his name with ground breaking work on Leontief
cost functions, was commissioned to produce today’s report as part of
the review into changes to the RPI.

He says he is focusing on changes to the RPI as the ONS has
control over this index whereas the CPI or HICP is controlled by
European legislation.

He comes up with several short term recommendations for reform of
the RPI.

The first is the scope of the RPI should be broadened “to cover the
expenditures of all households in the UK on consumer goods and

The second is the Carli index has “an upward bias” on the RPI, and
technical arguments people have put forward in favour of it “cannot be
used to justify (its) use.”

The third is “fashion goods should not be used as priced items in
the current RPI” and should instead be treated as “strongly seasonal

The Jevons method, which he suggests as one possible replacement,
relies on the geometric mean for prices, which entails that when prices
are dispersed it will be lower than the arithmetic mean employed in the

The Diewert report informs the consultation process on RPI reform,
which is due to end November 30.

The Consumer Price Advisory Council, which includes Bank of England
Monetary Policy Committee member Martin Weale among its members, will
make the final recommendation on RPI changes and the BOE will have to
rule on any potentially adverse impact on holders of index linked gilts.
If so, the final decision on whether to proceed with the recommendation
will lie with the UK Treasury.

Philip Rush, economist at Nomura, points out that buyers of index
linked gilts purchased them on the assumption that formula
effects meant that the RPI would differ from the CPI, and usually run
above it.

Removing the formula effect in full would have a significant
detrimental impact on these bondholders, whatever the pure statistical
arguments in favour of one methodology or another in calculating the two

“It’s statistics over common sense,” he says.

Ultimately if all the formula effects were removed the RPI and CPI
would be the same and “then what is the point (of two indices)?”, Rush

–London bureau: +4420 7862 7491; email: drobinson@marketnews.com