Measuring prices and inflation is central to almost every issue in economics: growth and
productivity; government taxes and benefits; budget deficits and debt; monetary policy; wage
setting; real financial returns; poverty rates; and the regulation of privatised firms. Yet measuring
the 'rate of inflation' - even defining what we should mean by that term - remains one of the
trickiest and oldest problems in economics. In the latest issue of the Economic Journal, Laura
Blow and Ian Crawford of the Institute for Fiscal Studies look at the question: how well does the
Retail Prices Index (RPI) measure the true increase in the cost of living?
There has been growing international interest in the question of how well official consumer price
indices actually perform in measuring the true increase in the cost of living. In the United States,
this was given additional impetus by Federal Reserve Chairman Alan Greenspan's calculation that
if the US Consumer Price Index (CPI) was found to have a bias of 1% per annum, then revising
index-linked government expenditure in line with this would save $55 billion after five years.
An expert committee was appointed to investigate the extent of bias in the CPI and their final
analysis - the Boskin Report - concluded that there was indeed an upward bias, of 1.1% per year.
This seems small but when compounded over time, it means, for example, that after twelve years,
the US national debt is increased by $1 trillion. If the RPI were found to be biased upwards, then
the implications for the UK, one of the world's leading issuers of index-linked government bonds,
would be no less dramatic.
The RPI formula, like many official consumer price indices, holds the basket of goods fixed for a
period of time as prices change. During this period, however, the basket of goods that consumers
actually buy may change, partly in response to these price changes. Fixed-basket measures do not
account for this. This usually means that there is an upward bias in the fixed-weight price index
compared to the true change in the cost of living. Blow and Crawford's study looks at the
performance of the RPI formula over the period 1976-97 and finds that:
l There is no automatic presumption that the direction of bias in the RPI formula is either
upward or downward.
l In practice, the (non-housing) RPI formula has overstated the true increase in the cost of
living over the entire period by between 1% and 3%. This implies that the average annual
rate of inflation as measured by the RPI formula over this period was between 0.06
percentage points and 0.15 percentage points above the true increase in the cost of living.
l The (non-housing) RPI formula clearly overstated the true annual rate of inflation in all but
three years of the period.
Notes for Editors: 'The Cost of Living with the RPI: Substitution Bias in the UK Retail Price
Index' by Laura Blow and Ian Crawford is published in the June 2001 issue of the Economic
Journal. The authors are at the Institute for Fiscal Studies (IFS). Their work was funded by the
Leverhulme Trust, the Economic and Social Research Council (ESRC) and the Bank of England.
For Further Information: contact Ian Crawford on 020-7291-4836 (email: ian_c@ifs.org.uk); or
RES Media Assistant Niall Flynn on 020-7878-2919 (email: nflynn@cepr.org).