Conditional and common data scores will have to be included in scheme returns from next year. Michael Klimes looks at what trustees need to do and if TPR's requirement goes far enough
Poor scheme data can result in money not being paid out to members on time and in the right way.
The Pensions Regulator (TPR) is concerned that too many trustees are not giving enough attention to what it sees as the two most important types of data - conditional and common data.
Common data is used to identify scheme members while conditional data is scheme-specific data to calculate member benefits. According to the watchdog's 2016 survey, 30% of members are in schemes where conditional data is not measured. Larger schemes are significantly more likely to have measured their data. Administrators and trustees are engaging with common data but are less engaged on conditional.
A tougher approach
These figures explain why the watchdog is now taking a much firmer approach. Conditional and common must now be included in scheme returns for defined benefit schemes from January 2018 and defined contribution plans from June 2018.
KGC Associates director Kim Gubler explains there is not a problem with trustees looking at both common and conditional data; the fact is a number of them simply do not examine it. "Often these are trustees of very small schemes. Our research and TPR's research shows these schemes have less engagement with their administrator and a smaller governance budget.
"So you have the combined impact of potentially poorer data and lack of connection between trustees and administrator. Making it mandatory to include the scores in the return is a strategy to change trustees' behaviour.
For Pension Administration Standards Association (PASA) chairwoman Margaret Snowdon TPR's robust approach is the right one.
"Voluntary guidance on data standards has resulted in little improvement over the last few years so having to report on record keeping as part of the scheme return is inevitable and will at least result in more schemes taking data seriously.
"Many schemes are unwilling to pay to clean up data as they don't see the link between good data and good governance and don't see that good data is a way of controlling risk. Administrators are sometimes reluctant to raise the issue in case they are blamed for the state of the data."
While many welcome the regulator's more muscular approach, will it actually improve the performance of trustees and is it enough?
Is it enough?
Even if trustees do develop a healthier attitude towards common and conditional data, there are concerns these will not improve data to the necessary standard to be truly effective.
Trafalgar House senior business analyst Claire Montgomery says there are two issues with TPR's new requirement:
"Firstly, conditional data is a gap analysis - it tells you whether something you expect to be in a field is there or not, the same goes for common data. Determining whether the data in that field is accurate is a whole different ball game.
"The other problem with ‘conditional data' is that it's a catch-all term for solutions that vary significantly across the market. There are some specialist providers who have highly sophisticated analysis and interrogation tools who can bring a great deal of insight and quality analysis to data as part of their ‘conditional data' analysis suite."
While these conditional data solutions cover a gap analysis in a few basic fields, they are not comprehensive and could lull trustees into a false sense of security, she argues.
She adds: "The danger is that trustees and pension managers see conditional data tests being run, with clean results, and assume they'll get that same result from anyone else who performs these tests. That's simply not the case and just like the quality of administration varies across the market, so does the quality of conditional data testing."
The shortcoming of such testing can be exposed at critical junctures in a scheme's history, Montgomery observes.
"Trustees are often surprised when they get to a buyout or transition administration services to a new provider, because a whole host of new data issues are found when much more detailed benefit assurance work takes place, even though they have previously completed a conditional data analysis."
The way to avoid such scenarios is to do a benefit assurance framework. "Benefit assurance projects take a much more detailed look at benefits through reconstructing liabilities on a first principles basis," Montgomery explains.
The downsides are benefit assurance projects are more expensive than conditional analysis reports, are time-intensive and require a far higher level of technical insight and expertise to correct data.
However, Montgomery says it "pays dividends in the long-run" and for maximum value "should be tied into a full legal benefit review to make sure nothing has been missed in the execution and application of historic rule changes."
"It also equips trustees and sponsors to undertake de-risking and eventual buyout negotiations with greater confidence and speed," she adds.
Capital Cranfield professional trustee Allan Course points out that if trustees have no reason to believe there is a problem with their data, then why spend money trying to check it if they believe it will be a waste of time.
Trustees might know areas where data gaps are most likely to occur such as people who work part-time, strange scheme jargon and special benefit guarantees. But if these are not on the records because they are lost or the trustees do not know where to look for them, solutions are limited.
"The company may have given all the information to the administrator, which on the surface looks complete but is not," he says.
"The danger is you don't know what data could be missing. Usually it is not immediately apparent. What do you question and how do you know what data is missing?"
Under these conditions, trustees must rely on a person with knowledge of the scheme sponsor. Alternatively someone must trawl through company records in the hope they will unlock the nugget of information to help pay benefits correctly.
"You hope for the best in all cases but there are three types of solutions here," says Course.
"The first is to question someone who was at the company during the past acquisition and re-organisation of the scheme. This is based on a person's memory.
"The second way to do it is hire a consultant with a checklist who questions the same person about every obscure word or special member guarantee they may have come across at their time in the company. You try and jog their memory and see what they tell you.
"The final way to do it is the most labour intensive. Here someone will go into the company archives and check all the literature. Although, the reality is the missing information you want might not even be there."
Course suggests TPR could boost data standards if it could have specific topics trustees could focus on for specific periods of time. "Perhaps it could be cyclical and take place over a five-year period."
For its part, the regulator is keen to get trustees to do better in their record-keeping duties. A spokesperson said : "Record-keeping is a vital part of running a scheme and failure to maintain complete and accurate records can affect a scheme's ability to carry out basic functions, lead to increased scheme costs and in some cases put members' benefits at risk.
"Record-keeping is not always seen as a priority by trustees, and they do not engage with their administrators accordingly - a quarter of administrators (23%), responding to a survey we conducted last year, felt that trustees treated record-keeping and administration as a low to middling priority.
"This situation is not good enough. We've made clear what our expectations are and from 2018 we will ask schemes to report on record-keeping in their scheme return
"We will provide further clarity on the content of scheme return later in the year."
Barnett Waddingham administration manager Ben Clacker thinks trustees should go beyond the regulator's new requirements.
"They need to go further than TPR standards in my opinion because even the conditional data will check an item in a certain field but will not check if that item is correct or not. That is the issue. When we inherit a new scheme, we go beyond TPR's requirements for data checks."
With scheme data an area of intensifying activity, trustees need to think long and hard about their approach to record-keeping.
PMI president Lesley Alexander and the institute's immediate past-president Lesley Carline talk about the challenges of Covid-19 and the opportunities and challenges the industry faces in the future.
The Pensions Administration Standards Association (PASA) has announced global consultant Deloitte as its expert knowledge provider for data.
This week’s top stories included further support for an overhaul of the pension tax regime, while the Treasury confirmed the Retail Prices Index will be reformed by 2030.
Tim Shepherd and Beth Brown look at the legal implications of working from home and how pension professionals can mitigate the risks.
Nearly every trustee is confident of the next stage in their scheme’s strategy, despite almost an equal number being forced to consider replacing plans within the prior 12 months, according to research by Barnett Waddingham.