Lynn Strongin Dodds explores why work still needs to be done to improve data, despite the regulator’s focus on improving record keeping

Data management may be a major theme in the investment world, but The Pensions Regulator’s (TPR) call to improve record keeping three years ago has not been a resounding success. Progress has been made, but there is still much more work to be done.

As data and systems specialist ITM’s director John Broker puts it: “The main issue is that poor data equals the wrong payments, which is why TPAS (The Pension Advisory Service) has been getting an increasing number of complaints year on year. The TPR issued a ‘guidance’ in 2010 that set out a clear basis for measuring data and had a sensible time frame to achieve this - 31 December 2012 - but it seems that it has been largely ignored.”

Pension actuaries and administration firm Spence & Partners’ director Neil Copeland agrees, adding: “Investment strategies and funding calculations get more attention but I think that TPR is correct in its view that trustees should be taking data more seriously. This is because flawed data will produce a flawed pension strategy. There has been a lot of engagement on common data but the bar is low although, even here, a number of schemes have not met the standards. Addressing conditional data is more challenging, but that shouldn’t be an excuse not to try.”

Common data is easier to unearth because it includes 11 categories such as name, date of birth and national insurance number information. However, although the regulator’s latest annual survey revealed that over 9.2 million trust-based members were in schemes which reached a 95 per cent score, this was not the case for around 2.3 million members. The common data in their plans were either not measured or fell significantly short of the target.

Punter Southall principal Gillian Graham also notes the results may be misleading because administrators employ different methods of crunching the numbers. “The problem I have picked up is that there seems to be some confusion about how the scores for the 11 tests are presented. For example, let’s assume that there are 100 members and when the tests are run, they pass all of the tests except that addresses are missing for 20 members. The correct score would be 80 per cent as there would be 80 members with no failures. However, some administrators would give this scheme a score of 98 per cent because they would score by test rather than by member, and in this case there would be 1,080 passed out of the total 1,100 tests. In other words, reporting results by test means that the results look much better.”

The issues are even more complicated when it comes to measuring conditional data that is bespoke to a scheme and features more detailed data such as contribution history, date of leaving and other items needed to accurately calculate a member’s benefit. The TPR study showed that this is not top of the agenda, particularly for larger trust-based schemes, where, of the 42 per cent that have not generated a conditional data score, 29 per cent said it was not an urgent matter. Moreover, half of small trust-based schemes who do not quantify this score were not aware of the requirement to do so.

“One of the main reasons for the problems is the evolution of the IT industry,” says Pensions Insurance Corporation chief admin officer Matt Gore. “As administrators migrated over to new systems, things got lost and corrupted during the journey. Also, the recent trend to outsourcing has had an impact. It is costly to fix data and covering funding levels at the moment is a higher priority although I think the regulator will continue to push harder for quality data.”

Looking ahead, few market participants expect to see an imminent change. “The regulator should focus more on conditional data and give it more teeth,” says Pensions Administration Standards Association (PASA) chair and JLT director Margaret Snowdon. “Unlike common data, they have not set out any timelines. However, there is little appetite to fine schemes. This is mainly because this would affect the smaller schemes more and it could impact members’ benefits.”

Written by Lynn Strongin Dodds, a freelance journalist

Share Story: