What is the Self-Administered Pension Scheme (SAPS) Mortality Investigation?

Currently pension schemes in the UK use mortality tables based on population data or data collected from insurance companies. These tables may not be suitable for self-administered schemes.

In 1997 the Pensions Board of the Faculty and Institute of Actuaries and the CMI began considering the feasibility of research into this mortality experience and it was decided to run a pilot investigation on a small subset of self-administered schemes to see how the mortality compared to that of insured pensioners.

The results of this investigation were published in CMIR 20 in 2001 and concluded amongst other things that the experience investigated was different to that underlying the insured pensioner investigation.

In 2002 further discussions were held and a decision to begin a more wide-ranging and substantial investigation into the mortality of self-administered schemes was taken.

When did the investigation commence?

Data collection began in early 2003.

Who contributes data to the investigation?

Eleven firms have contributed so far. These are mainly the larger consultancies, although smaller consultancies have also supplied data. One insurance company has also submitted the data for its pension scheme. Data is collected for schemes with over 500 current pensioners.

How much data has been collected so far?

Over 250 schemes have so far been submitted. The schemes are from across the board of industrial groups (as shown in the Financial Times) and should therefore give results that reflect the whole of the UK self-administered pension schemes population.

What results are produced from the investigation?

Two types of results are produced. Initially, comparisons of actual deaths with the deaths expected according to certain standard tables are reported. These analyses are done for each scheme and reported to the consultancy who supplied the data. Comparisons for Males and Females, by Lives and Amounts, and for different Pensioner Types are produced.

Aggregate results are then produced, analysing across all schemes, which are normally published within Working Papers.

It is hoped to produce standard tables of mortality based on all the data collected over a number of years. In the longer term, information about mortality trends will also become available.

When should data be submitted?

In most cases data submissions are made soon after the scheme valuation, with the data coming straight from the valuation data. This will usually be every three years, although schemes having annual valuations can submit each year’s data separately.

In what format should data be submitted?

The coding guide contains all the information required. Click here for the latest coding guide. If you have any queries about data submission, please contact us on 020 7776 3820 or email self-admin@cmib.org.uk.

How are data problems handled by the CMI?

To keep the costs of the investigation down, the data must be provided in the standard form shown in the data specification. The CMI makes a limited number of checks on the data. If an error condition on a record can be corrected in some predetermined way (e.g. a missing amount of pension at the date of death) the system automatically updates the record (although in these cases it is preferable for the submitting firm to estimate these). If this cannot be done then the record is rejected. The analysis is then performed on those records not rejected. For the combined analyses, a whole scheme is removed if the number of the records that are rejected exceeds a predetermined threshold. Along with individual result reports an error report will be produced, detailing those records that have been altered or rejected. The scheme actuary may then resubmit the whole scheme with changes if required. Two of the fields, the amounts of pension at the start and at the end of the investigation period, are voluntary. If these are not supplied the CMI will make assumptions as to what they might be and will show this in the error report.

The summary reports allow each consultancy to inspect the results for their schemes for reasonableness. For example, the number of lives covered and the number of deaths experienced are shown. The CMI does not inspect the results in respect of each scheme but relies on those submitting data to do so.

Results show Actual experience versus Expected. How is Expected calculated?

The expected experience is calculated by applying the exposure at each age with a standard table of mortality. There are two bases currently used for these comparisons: the first basis is PA(90) for amounts analyses and a(90) for lives analyses (based on the 1967-1970 investigation). The second basis is the 92 series for both lives and amounts analyses (based on the 1991-1994 investigation). Note that these standard tables are derived from data on insured business.

If pension review dates change from year to year or there is more than one review date, what should the scheme actuary do?

Please make an estimate of the average review date applicable in the inter-valuation period covered by the data. If the pension review date field is left blank then the CMI will assume a mid-year pension review date.

What type of pensioners should be included in the data?

The data should contain records in respect of retired members or pensions paid to widows or widowers of scheme members. Data in respect of child dependants should not be included.

Should data be submitted in respect of deferred and active members of the pension scheme?

No. Data should only be submitted for current pensioners.

Does the CMI collect data for schemes with less than 500 members?

At this stage only large schemes are requested to submit data. This number was based on an analysis of the expected number of schemes and the cost of the whole investigation. Including smaller schemes would add to the cost without adding significantly to the validity of the data collected.

Where can I find results?

Results for all schemes combined have been published in 3 Working Papers to data. Please click here to access these papers.