A group of leading UK retail industry pension schemes are set to link up with the Continuous Mortality Investigation (CMI) unit to produce life expectancy figures specifically relating to this sector - the first time such a task has been undertaken.
CMI tables are compiled from 80 years of pooled life expectancy data from insurers annuity portfolios. With customers of such products tending to be wealthier and better educated - two key indicators of life expectancy - than the majority of retail workers, pension providers, such as Sainsbury, for the latter group are concerned over the issue of basis risk in mortality figures.
"There's a problem with life office data," says Chris Armitage, director of pension investments for the £4.3 billion Sainbury's pension scheme. "Typically these annuitants are company director-level people with life policies. The members of Sainsbury's scheme are quite different."
Up to now, pension actuaries have attempted to update standard tables such as PMA92 using cohort adjustments that reflect improvements in life expectancy, which increase estimated liabilities. Confirming that Sainsbury's had used the medium cohort adjustment in its last triennial valuation, Armitage raised the possibility that the industry initiative could result in a downward revision of liabilities in the future. The revisions are potentially significant because although it is closed to new members, Sainsbury's is a relatively immature scheme with a duration of 28 years.
Three years ago, the CMI group set up the Self Administered Pension Scheme Survey (SAPSS) to look at the issue of differing mortality projections in varying industrial sectors. But according to Nigel Bodie, a SAPSS committee member, this is the first attempt to extrapolate data that can be used by individual schemes to compare their mortality experience with their peers.
Work will start in January 2007, but with a number of schemes in the sector currently conducting their tri-annual valuations, Bodie said he didn't expect all the data to be in until the end of March, with the results expected in summer 2007. He said it was likely that this approach to mortality investigation could be broadened to include other sectors in the future. "This approach could be taken by other industries, but it would only be of value if they are as homogenous in terms of member profile as the retail industry," he said.