Why data scoring is the key to a new era of scheme administration

18 December 2024

With the support of The Pensions Regulator (TPR), the Pensions Administration Standard Association’s (PASA) Data Working Group has launched a discussion document on data scoring – Decoding its history, significance, challenges and TPR’s expectation.

The report discusses recent regulatory initiatives aimed at improving data quality. And with no blame attached, it describes the current shortcomings of data scoring and makes recommendations for improvement.

But it’s more than a discussion document. In the appendices, PASA has helpfully provided lists of data items that schemes should focus on when testing data for:

1. All members
2. In-service members
3. Deferred members
4. Pensioners and dependants.

This is an important step towards aligning data scoring across the pensions industry and improving transparency of scheme data quality.

The timing is critical. Our new digital world makes heavy demands on the data we hold. We no longer live in a world of mainframe computers, print-based archives and typing pools. We’ve got automated systems which run on structured data.

Clear transparency of data scores in place will transform the scope for delivering the enhanced member experience and efficiencies that the new digital world offers, with data-driven technologies such as AI.

But there’s a long road ahead.

Where are we now on data scoring?

At present, there’s a worrying mismatch in approaches to data checking, especially where scheme-specific data is concerned. And not all trustees add data scores to their scheme returns on a regular basis.

Consistent and repeatable data scoring can only be achieved if administration data is digitally accessible, stored in one location and recorded consistently for all members. Right now, only a small minority of schemes are in that position. In a former paper-oriented industry like ours, old habits die hard.

In the main, common data – the information all schemes hold – is relatively problem-free. That said, requirements around common data are in flux. First name initials of members are a great example of data that has previously been acceptable but is no longer adequate. A letter to a member can be addressed to Mrs C Fuller, rather than Mrs Claire Fuller, for example.

In the digital world, this isn’t enough. With pensions dashboards, data matching isn’t possible without a member’s full name. Trustees need to adopt a digital mindset, and hold data that is fit for the digital world.

And then there’s data accuracy. Up to now, we’ve assessed data on the basis of presence – checking that we hold data in the date of birth field, for example. But we haven’t measured the accuracy of the data itself. First Actuarial contracts a third party verification service to check member addresses every year. Across the industry, we need to do more of this.

An important change is that trustees must consider their long-term outlook of their scheme, and not just business-as usual activities and standalone projects such as GMP rectification. Schemes should start to think more broadly about future demands on data, rather than taking a piecemeal approach.

We’re going to be asking so much more of scheme data than we did before, and increased stringency and consistency come with that.

Why are schemes in this position?

Up to now, there has been no real regulatory enforcement in this important area. Without that, trustees have focused on other priority areas such as their investments.

That’s not to say that there has been no regulatory output. As early as 2010, TPR set out its expectations for assessing common and scheme-specific data. Only a small number of schemes took action at that point. When TPR updated its guidance in 2016, we saw more schemes make changes to their data scoring. But it’s been a slow process.

And that’s why we felt, as an industry group, that closer focus and increased support were needed.

What are TPR’s expectations?

The Pensions Regulator has launched its data quality initiative. Schemes now have to submit data scores on an annual basis, rather than once every three years as they’ve done up to now.

TPR has contacted those schemes that have not been providing scores as part of their scheme returns, making its expectations around data scoring abundantly clear.

TPR is planning to analyse the scores from next year’s scheme returns, to get a benchmark of data quality and scoring. It’s important to emphasise that schemes shouldn’t worry about submitting a low data quality score. Low scores are what the industry fully expects right now.

What matters at this stage is that schemes should be actively looking at their data. It’s when 2026 scheme returns come in that TPR will want to see improvements in scores.

What administrators should be doing

Now is the time for scheme administrators to shine. Data underpins everything that happens in the pensions industry, and it needs to take centre stage.

The good news is that administrators are well-equipped to meet the challenge. Most if not all administrators have great data skills. The trick is to start looking at data in a different way.

We tend to focus on data at the level of individual members, as we support them through their journey. We need to expand our outlook – to see the bigger picture, and to consider both the present and the future. We already do that to an extent with work such as the regular data overhauling we carry out. We just need to widen our data view a little bit more.

In the immediate term, administrators should review their data scoring analysis and provide assurance to trustees that a full data analysis has been carried out. They should state clearly what has been analysed and how the tests have been performed. Data scoring should be consistent and comparable.

The better our data gets, the more interesting and useful scheme administration can become. We need to get our data into a fit shape for the future that is already here.

We’re here to help schemes

The mood music from both PASA and TPR is one of support rather than judgement.

Yes, the report talks about measuring and comparing schemes. This can conjure up an image of a pension scheme data league table. But it’s about setting a benchmark and then helping schemes as they progress with data quality and scoring improvements.

First Actuarial’s administration team has extensive experience of common and scheme-specific data scoring. We’ll be speaking with all our clients, to help them meet TPR’s requirements, put together a data management plan, and future-proof their data.

Data needn’t be boring. This is an exciting time for everyone involved in pensions data, as exciting new opportunities open up in scheme administration and the member journey.