ONS identifies error in RPI calculation
9 Aug 2019
The Office for National Statistics (ONS) has admitted finding an error in the way the Retail Prices Index (RPI) calculation was processed in 2019, casting further doubt on the validity of the measure used to calculate payments on billions of pounds’ worth of inflation-linked government bonds
9 Aug 2019
The ONS says the mistake was caused by an issue with the 2017 to 2018 living costs and food survey (LCF) dataset, which is used to produce the weights underpinning the RPI.
Indicative estimates show that if the corrected LCF dataset had been used to calculate the 2019 RPI weights, it would have led to an upward revision of 0.1 percentage points to the published RPI annual growth rate in March 2019, from 2.4% as currently published to 2.5%.
There would also have been a corresponding downward revision of 0.1 percentage points to the June 2019 rate, from 2.9% as currently published to 2.8%. No other month’s annual growth rates have been affected.
Using the corrected LCF data would also have led to an upward move of 0.1 index points to the March 2019 RPI, from 285.1 to 285.2, and a downward move of 0.1 index points in each of April (288.2 to 288.1), May (289.2 to 289.1) and June (289.6 to 289.5).
In line with the government’s revisions policy for consumer price inflation statistics, the RPI will not be revised. This error has been corrected from the July publication, which will be published on 14 August. The growth rates and index levels will be published as usual using the corrected weights.
However, the ONS says users should note that as previous months’ publications will be on the previous basis, there will be a discontinuity between July’s and June’s index levels.
The error does not have any impact on either the Consumer Prices Index including owner occupiers’ housing costs (CPIH) or the Consumer Prices Index (CPI), for which weights are calculated using more comprehensive national accounts data sources. These sources draw on a wider range of survey and administrative sources than just the LCF.
The ONS came under fire earlier this year over issues with calculating RPI. In January a House of Lords economic affairs committee report revealed a statistical error in RPI caused it to artificially increase by 0.3 percentage points in 2010, leading to a £1bn yearly windfall for index-linked gilt holders at the expense of consumers such as students and rail passengers.
This committee and the Treasury select committee then wrote to UK Statistics Authority (UKSA) urging it to seek consent from the Chancellor of the Exchequer to fix RPI. In 2013, the measure was judged to fall short of international standards, but continues to be published because about £400bn of government debt is inflation-protected using the RPI index.
Pat Sweet | 09-08-2019