Order, order, is your data in order?

by Colin Wheeler   •  

Data, in the context of pension schemes, has probably never been talked about as much as it is now. And there is one question that trustees keep getting asked - is your scheme’s data “in order”? But just what does this mean?

To those not immersed in the nuances of the subject, it may appear that data is data - how can it not be in order? When it comes to pension schemes there are a multitude of reasons as to why this is the case. For example, if we consider a typical member who joined a scheme in the 1980s, some or all of the following may have happened, which could cause their data to ‘decay’ over time:

  • employer is taken over
  • scheme administration is outsourced
  • scheme records went from paper to microfiche to computer based
  • scheme has passed through several third-party administrators
  • member left and became a deferred member
  • member retired
  • legislative changes caused benefits to need tranching-out
  • member transferred-in a previous company pension
  • scheme was merged into another run by the same or a different employer.

The list of reasons could go on and on.

During all the above, the member’s data will have been handled, giving rise to the possibility of it being changed, correctly or not.

Administrators, trustees and sponsoring employers all have a part to play in the data journey. Administrators should be reporting on data on a regular basis and, when we say data, it should be data which is meaningful in the context of buy out, so conditional or scheme specific data. Trustees and sponsors have a duty to be taking these reports seriously and taking the appropriate actions coming out of these reports. Over the last five years or so, many data reports have been produced with little or no thought given to their eventual use, and have merely ticked a box for trustees. This is not good enough.

High expectations

Data will need to be in order for a number of projects, including member derisking exercises, full calculation functionality and member online portals. In this blog, I want to discuss specifically data requirements for buy out.

When trustees go to market to insure member benefits via a buy-out, or a buy-in for that matter, there is an expectation placed upon them by insurers that they will have taken the necessary steps to get their data in order. In turn, trustees will expect their administrators to have data ready to share with insurers when quotations are sought. It is this engagement between trustees, scheme sponsor and administrators which is key to data readiness, and which needs to be in the project plan and discussed at an early stage.

When reviewing data in preparation for buy out, it is necessary to think about the insurer requirements, and the issues which will have the biggest bearing on their pricing, and indeed whether they will want to quote at all. The capacity in the market is limited and insurers can afford to be selective over which schemes they want to transact with. A scheme which can demonstrate that it has taken the time to prepare for going to market is more likely to appeal to insurers than one which hasn’t. Schemes which have invested in data preparation will see this reflected in the price compared to a scheme with incomplete data

For a relatively mature defined benefit scheme, the typical data issues which are likely to need addressing could include:

  • Pensions payable to spouses or dependants on the death of a member. This is particularly the case with pensioners who commuted pension for cash, but the pension on death is based on the pre-commuted pension.
  • Tranching of benefits, particularly for deferred members where different revaluation rates will apply to different tranches, and indeed some tranches will have different minimum retirement ages.
  • Treatment of transferred-in benefits for deferred members. This assumes that the transfer-in can be identified in the first place!
  • Data on marital status of members, and for those married, their spouse’s dates of birth.

It is not just a scheme’s core data that insurers will expect to receive. Other data not needed for day-to-day administration, such as mortality experience and marital status, can help insurers when valuing liabilities. The more data a scheme provides, the fewer assumptions an insurer will be required to make, which will allow it to price with a degree of accuracy.

It is also worth noting that many schemes will have members with “special benefits”, which have been documented in individual member letters or augmentations agreed with the sponsoring employer. In such cases the trustees must ensure that all instances of those benefits are identified and reflected in the data on which insurers are asked to price.


This is an area that insurers will look closely at, to understand exactly how the scheme equalised retirement ages, and how this has been reflected in data. They will seek the usual legal evidence, but when it comes to data, they will expect to see the respective “Barber” periods of service tranched out separately, with clear guidance to show what retirement age applies to each.

For schemes that tackled equalisation by awarding additional service, or some other form of compensation, this needs to be clearly labelled in the data. We may be 30 years on from the original Barber judgement but it is still proving to be a thorn in the side of scheme administrators. And Guaranteed Minimum Pensions (GMPs), specifically inequalities arising from GMPs, are rearing their ugly head, but that’s another matter…

It’s not only data that is needed

As well as data, a scheme going to market will need to provide a comprehensive, and legally endorsed benefit specification. The insurer will use this to interpret the data and to test that the scheme is being administered in line with the rules. It is not uncommon for schemes to have specific practices which have evolved over the years and which are applied ahead of the scheme rules. So, if the scheme’s legal advisers are preparing a specification, it is crucial that the administrators review this to ensure that it represents what happens in practice. It then becomes a matter for agreement between the insurer and trustees as to the basis on which policies are set up. If policies are not correctly set up then the trustees may find that they are not fully discharged in respect of the members bought-out. 

So, there you have it, getting data in order is not a task to be taken lightly, or left too late. When it comes to buy out of benefits, preparation is the order of the day.

Further reading

Spence Quarterly Update Q4 2020

by Andrew Kerrin   •  

The threat of inflation

by Brendan McLean   •  

Government spending in response to Covid-19

by James Sweetnam   •  

More Insights?