In order to assess long-term DB liabilities, some estimate of inflation is crucial. In the UK, the RPI used to matter, not the CPI or CPIH. With effect from April 2011, the public service pension scheme increases were switched from RPI to CPI, which some private sector pension schemes were able to follow.
Not only are preserved benefits normally partially linked to inflation over the whole deferment period, but also pensions in payment are partially linked to inflation on a year-by-year basis. Some pension schemes grant full RPI coverage whereas others have a 5% limit (few schemes appear to have adopted the 2.5% limit permitted since 2005). Even for a lower limit, one still needs some idea of how much the full increase would be and how much to allow as a margin.
Although UK inflation has hardly been above 5% in any year since the early 1990s, that has simply not always been the case, with some of us recalling the 1970s. At the time of writing, even CPIH has increased by 6.2% in the last year and is expected to increase further. Why should we assume that it must remain under control? Yes, the Bank of England has a remit to keep inflation under control but that is currently based upon the CPI rather than the RPI and they are regarded as having failed in that mission. Increases comparison charts for and between CPI, CPIH and RPI are shown, linked from RPI alone (since 1953), RPI_v_CPI (since 1975) and CPIH_v_CPI (since 1989).
We're worried that liabilities are being hugely overstated, by the order of at least 25%, possibly even higher. While we don't claim to know the future, some estimate must be made and we are worried that current estimates are just plain excessive. Oh, no, we don't believe there is a holy grail. Further, we are surprised that major UK actuarial consultancies can assume a long-term average increase difference as high as 1% pa between CPI and RPI. While we have seen that in single years, we have rarely seen it over periods as long as 10 years (15 out of 46 between 1975 and 2021) and never over periods as long as 15 years.
Comparing RPI with CPI over 1 year, the average has been 0.73%, with a standard deviation of 1.02%. As it occurs at the 60th percentile, it is not unreasonable to suggest that the expected difference could be as high as 1%.
However, comparing RPI with CPI over 15 years, the average has been 0.72% (virtually the same as over 1 year), with a standard deviation of 0.10%. A value of 1% would be 2.84 standard deviations away from the mean, which corresponds to a likelihood of 1 in 200, namely highly unlikely. Hence, 1% really cannot be taken as a best estimate.