Comparing household inflation experiences measured by the CPI and RPI

In October 2012, the ONS announced a consultation on whether the statistical methods used to calculate the Retail Prices Index (RPI) should be changed to bring them closer in line to those used in the Consumer Prices Index (CPI). Previous IFS work has looked at how inflation rates varied across different households, using survey data on household expenditure to calculate RPI-based measures of household-specific inflation. This paper analyses whether CPI-based measures give similar results and the reasons behind any differences. In doing so, we investigate whether proposed methodological changes to the RPI would have changed our previous results on the difference in inflation rates across household groups had they been implemented before. We find that, after stripping out housing costs, there are only small differences between RPI and CPI-based measures of the gaps between high and low income households and pensioner and non-pensioner households. This suggests that the 'formula effect' difference between the two indices doesn't systematically affect the goods typically consumed by either pensioners or low income households more than the goods typically consumed by non-pensioners and high income households.