I first became interested in Inflation measurement some years ago when GFK bid (unsuccessfully!) for the data collection work that inputs into the UK RPI.
At that time the prices for each of the products and services in the basket of goods were collected by a team of market research Interviewers and the request for proposals as provided by the ONS had requested innovative ideas. Our contribution was to suggest that commercially available point-of-sale data could be used to significant benefit (in terms of accuracy) in several areas: specifically food products (where statistically reliable data was at that time available from the Neilsen Food Panel) and technology products (TV’s, mobile phones etc) (then available from the GfK Retail and Technology Panel). Both of these categories were particularly problematic since products and specifications changed rapidly and maintaining comparability was a major headache. Indeed several Statistical Offices in more innovative countries were already experimenting at that time with integrating commercial data into their indices. However, despite the request for ideas in the RfP, there was in reality curiously little interest in innovation in the UK.
And I’ve observed many times since how difficult it is for clients to change methodology on large scale continuous surveys – fear of changing the data keeps them locked into designs which may have been state-of-the-art when they were first devised, but have become increasingly compromised by changing behaviours and potential new approaches over time.
I am therefore sympathetic to the underlying desire to create a more modern process for the RPI and conscious of the efficiency to be gained by linking it directly to the international standard macro-economic inflation measure, CPI. However that does not make it the right thing to do.
As is the case in many other countries we should continue to use a measure that genuinely reflects inflation as experienced by consumers. That means using a methodology more consistent with that informing the RPI, not the CPI (or even the CPIH).
In turn that means using a process of continuing development, experimenting with new methods in parallel with the old and comparing results. Such procedures have been common in the commercial world to maintain consistency of measurement in a changing world. Surely we could expect something similar for a measure as important as the RPI?