Large tracts of data held by defined benefit schemes in the UK are held using state of the art storage techniques. Unfortunately, schemes hold data going back over several decades, during which time the ‘latest technology’ was in fact cards, paper and microfiche.
While some pension schemes have made serious efforts to modernise and transfer their data into a digital format, many schemes have yet to do so.
Our interviews with 88 pension professionals, conducted in December 2013, found that 17 per cent think the quality of information held by DB schemes is ‘poor’ or ‘very poor’. Only 1.3 per cent would describe it as ‘excellent’. Interviews with 163 pension professionals in June-July 2013 showed that only 51 per cent of pension professionals believe more than half of DB pension data is stored in digital format, although 78.3 per cent do believe this will be the case in five years’ time, and 16 per cent believe all data will be digitised.
Many of the issues and problems around storing data in a variety of formats emerged during our recent EDM roundtable discussion.
The panel included pension scheme administrators, a director from a leading audit firm, industry consultants and a leading academic and trustee. The main message to emerge was that DB schemes need to modernise their information storage as pressures mount on them to de-risk and meet new regulatory requirements. Furthermore, these pressures will increase over the next five years, and 2016 will be particularly crucial because of data management issues inherent in the government’s implementation of the single tier state pension scheme in April 2016.
DB schemes’ desire to de-risk is well-documented: falling gilt yields, longer life expectancies and inadequate investment returns, especially from equities over the last decade or so, have all put pressure on schemes as their liabilities have ballooned.
Enhancing the quality of data makes decision-making easier. For example, holding data on a variety of mediums can make it hard to quantify the actual liabilities of a scheme, which can scupper M&A activity among scheme-sponsoring companies. Unknown liabilities also deter insurers offering buyout and buy-in deals. Insurers offering these do not wish to take any potential risks with liabilities that they cannot quantify. John Lewis Partnership Pensions Trust head of trustee services Martin Mannion told the roundtable that an insurer will insure any amount of uncertainty. “But when a scheme goes to an insurer and says I only have 50 per cent of my data the insurer will say fine, here is the premium for the bit you know and here is the premium for the bit you do not. And the difference is eye watering.”
Grant Thornton head of pensions audit Fiona Baldwin told the roundtable panel that the additional premium added to buyouts due to poor data can be up to 10 per cent and sometimes considerably more, so ensuring data is as clean as possible has to be a priority for trustees who are considering de-risking.
Cass Business School visiting fellow, and Merchant Navy Ratings Pension Fund investment committee chair, Chris Wagstaff, identified a pattern in the de-risking process that schemes are following. He said the vast majority of DB schemes have so far engineered some form of interest/inflation rate hedging programme before taking their de-risking programmes one step further and considering longevity risk, which is a more drawn out process and difficult to hedge.
However, with current yields at such low levels, many schemes, not least those that have failed to plan their de-risking ‘flight plan’ properly and have held back because of concerns about where interest rates are heading, now have to consider more complex interest rate hedging solutions that demand an advanced level of trustee investment governance. Consequently, there are a lot of schemes with bloated liabilities, often bigger than the market cap of the sponsoring company, that have missed a great opportunity to de-risk at a time when insurance rates have been quite favourable.
Wagstaff added: “It is important to differentiate between the data required for interest/inflation rate de-risking, for which there is plenty of data, and that for the next stage of de-risking, for which data cleansing is the imperative. As more schemes take out interest rate/inflation risk the longevity risk becomes more prominent, and to take that risk off the books efficiently you need clean data.”
Regulators are also increasing pressure on trustees to improve their data storage. The Pensions Regulator intends to review its record-keeping guidance for schemes this year following the passage of the current Pensions Bill and the outcome of a thematic review of 250 schemes. Its own research has shown that over two million members are in schemes where common data such as names, dates of birth and National Insurance numbers are not measured or significant improvements to record keeping are needed. Independent Trustee Services trust officer Janet Branagh said regulators are requiring schemes to “ensure the right data is held in the right way”.
The government’s announcement that it will bring forward implementation of the single-tier state pension from April 2017 to April 2016 has increased the data challenges for DB schemes and means they will have to move even quicker than they might originally have thought. From April 2016, HMRC will no longer track pension rights: it will be schemes’ responsibility to keep track of who their members are, while members will have to know which scheme holds their pension rights until they reach state pension age.
A key requirement by April 2016 will be that all DB schemes will have to hold accurate assessments of their members’ guaranteed minimum pension (GMP) rights so they can be compared with the National Insurance and PAYE service systems held by the government. From April 2016 HMRC will introduce an online channel through which schemes can receive GMP calculations on demand, both on an individual and bulk basis. Responses will be provided electronically.
Branagh pointed out: “Unless they (schemes) have already undergone a data cleansing exercise, this will potentially expose gaps in their data and significant resources will have to be spent collecting it, potentially increasing the administration costs.”
Ensign director of administration Michael Mann said the impending changes due in 2016 posed a risk for trustees, who will be responsible for schemes keeping accurate records of their members’ guaranteed minimum pensions. He said schemes with paper or microfiche records could end up paying inflated costs as administrators will have to sift through such records to calculate the member benefits payable. He said: “We often find that 10 per cent or 15 per cent of schemes have members they did not know they had when completing an historical GMP reconciliation.”
Despite the strong incentives to digitise data, there is some inertia preventing schemes from doing so. Baldwin commented: “I know many schemes that recognise there are issues they need to fix. Maybe that is what holds them back: they may realise they could be opening up a Pandora ’s Box. But that is a very short term view. It is better to do data cleansing if you can. The cleaner your data the more informed the decisions you can make.”
Having appropriate records is as fundamental and vitally important as having proper books and accounts. It is easier to solve the issue of incomplete or inaccurate data than many people realise in terms of doing a data cleanse and making sense of what is there. Many seem to think that because it is on microfiche, paper and cards going back 40 to 50 years then it is a major problem. But there are all sorts of techniques to get data in the right format. I cannot understand why the market is not progressing this more resolutely – but it is clear that events may well overtake those companies determined to delay.
Mark Jones is chief financial officer, EDM Group