Data is everything and everywhere, from generative artificial intelligence (AI) to supermarket loyalty cards.
So, it is inevitable that data – either a lack of it, too much of it or even the wrong kind – is proving a challenge to the UK’s pension risk transfer (PRT) market.
The latest Pensions Funding Survey from PwC, published at the end of October this year, highlighted growing concerns over data quality. A quarter of the trustees and managers, jointly responsible for £200bn of defined benefit (DB) pension scheme assets, said their number one concern was data preparation and third-party administrator capacity.
PwC says that the industry has worked on a ‘just in time’ basis, meaning data was often not checked until just before a member retired, and Gareth Henty, Head of Pensions at PwC UK, said the current climate meant this approach was no longer appropriate.
“The pensions industry is facing issues with schemes having their data ready for buy-out or risk transfer, along with getting data and systems ready for the introduction of pensions dashboards and completing Guaranteed Minimum Pension equalisation (GMPe) exercises,” he said.
Frankie Borrell, Head of BPA Origination at insurer Royal London, admitted data was proving a challenge to many of the 5,000 defined benefit UK pension schemes, plenty of whom have historically managed with member data that came with “varying degrees of inconsistencies and gaps”.
“Buy-out providers need a certain standard of data to meet their requirements, both in terms of accurately pricing the insurance risk and being able to go on to administer those policies for decades into the future,” he said.
“Over the past two years we’ve seen a much higher interest in buy-out policies from pension scheme trustees which in turn has led to a much greater strain on administrators preparing for ‘insurer ready’ member data. This strain has been compounded by various other member data initiatives, not least the need for the vast majority of these pension schemes to now implement GMP equalisation following the Lloyds judgement.”
The Pensions Regulator (TPR) is pragmatic about scheme data readiness but is insistent all schemes review their data at least once a year and put an immediate improvement plan in place to address any issues.
The TPR also highlights the key role of scheme administrators, which it states: “should be able to help you design an improvement plan or assess the one you currently have in place”.
“Having accurate and timely administration is the bedrock for any strategy that trustees and employers want to pursue,” Henty said.
“In addition, as trustees and sponsors will, for the first time, have to document their long-term objective in a Statement of Strategy, and as part of this due consideration should also be given to a scheme’s data strategy and the resource and capability needed to execute this over the period until the long-term funding objective is reached.”
Hymans Robertson’s paper Why data is the queen in a scheme’s end game was published in October and offered a five-stage approach to schemes wanting to ensure their data is fit for purpose.
“Every DB pension scheme will have its own journey but taking a holistic approach to data improvement and working towards an ‘accurate all the time’ data set will benefit the scheme’s progression towards their chosen endgame,” said Hymans Robertson Head of Digital Strategy, Scott Finnie.
“This will allow for a smoother transition and increased flexibility as all end game options can be explored with the knowledge that the data is accurate and correct.”
Hymans’ approach starts with a clear definition of the link between data and benefits. Stage two involves auditing and reviewing the data at regular intervals while the third stage is a data improvement plan; stage four is about executing the plan and providing regular updates to stakeholders and stage five means maintaining ongoing monitoring and seeing data as a “key integral part of a scheme rather than a one-off exercise”.
“Having correct data benefits DB schemes in several important ways,” Finnie added.
“It cuts administrative processes, reduces risk and enhances the member experience – all of which are invaluable to the smoothness of a scheme’s endgame journey. Investing in a considered, holistic data improvement plan will make the scheme more attractive to the market – as well as offering reduced risk, dashboard compliance, and improved member experience along the way.”
Julie Yates, Director of Pensions Administration at actuarial consultancy Cartwright said it was seeing several different logjams around data. For example, the documented benefit specification document could sometimes raise more questions than answers.
“Trustees need to check that the benefits being agreed with insurers actually agree with what the scheme rules set out,” she said.
“This could result in a review of scheme trust deed and rules and any deeds of amendment and carry out a benefit audit. Any mistakes need to be sorted out before going to market, as trustees won’t want to run the risk of securing the wrong benefits, and incurring ongoing liability after wind-up, because of poor data.”
Yates recommends ‘pre-work’ to establish gaps in data.
“The need for clean data goes far beyond the regulator’s current guidance and potentially even the ongoing administration records. Each insurer requires their own templates to be used and having an effective and open relationship with them, to ensure data is of a high quality and provided in a timely manner is essential.”
Yates said projects to validate such items as member’s marital status, dependants and addresses ensured there were no hidden surprises as part of any top-up premium that would be payable as a buy-out is completed.
“Starting the process with poor or incomplete data simply puts more risk on the table which could result in a large and an unexpected bill for the sponsor,” she said.
GMPe projects were also proving a challenge; trustees should be aware that GMPe projects are ideally completed before going to buy-in.
“GMPe work undertaken during the buy-in stage could delay schemes completing the transaction within the desired timeframe and what is required to complete these projects successfully,” said Yates.
A review of the required data that must be in place in order to perform accurate calculations needs to be undertaken as early as possible, she said. Schemes must also weigh up the cost of undertaking a GMPe data cleanse and completion of the GMPe project with the benefits of a smoother transaction and potentially a lower than otherwise insurance premium.
“Insurers, rightly, want data in their own templates and this can involve re-trancing of pension benefits. Preparation and understanding the requirements in advance will remove any last-minute obstacles to transact and will ensure the ongoing administration from buy-in to buy-out works smoothly,” said Yates.
“Neither the current administrator nor the insurer wish to start off on a bad footing, where data is not up to requirements.”
Getting their data ducks in a row is a scheme ideal, but insurers are realistic about the challenges the industry faces.
“As a newer entrant to this market, one of the key triage factors for us participating in a quote process is the quality of member data,” said Borrell.
“Whilst we don’t expect perfect data pre-transaction, we always seek to work with trustees and advisers that have taken the time to properly prepare their data so our pricing and onboarding processes will run smoothly.”
