Optimising the risk mitigation impact of data cleansing
It is a truth universally acknowledged that good data is important for pension schemes but I wonder if we all mean the same thing when we say that. Do we mean data or do we mean benefits? Are benefits data? The point of data cleansing is to get the benefits right. Is it possible to do that without legal input? Without legal input we are only doing half the job.
Some schemes are seeking legal input and my colleagues and I spend a lot of our time doing that kind of work but it’s mostly where buyout is on the horizon and the cost benefit equation looks a bit different. The pressure for all schemes to get the benefits right is increasing. The Dashboard will increase exposure and member queries. More will be expected of professional trustees.
The reluctance to find new problems is understandable, on everybody’s part. Perfection is not achievable. It’s a question of balance but turning a blind eye may not be the best commercial strategy in the long term. Advisers and administrators need to have some grown up conversations with trustees about what they mean by data cleansing and what are their objectives.
Real world facts
A distinction is often drawn between data and benefits. That makes sense if ‘data’ means personal information like date of birth, full name, gender, NI number, address, date of joining, date of leaving, date of retirement. These are facts. Where they are not present or implausible you can tell. And it’s worth doing some work to see if that information can be found.
Legally defined real world facts
Pensionable salary and pensionable service depend on two things:
- real world facts such as pay and bonuses, employment dates or periods of absence, and
- legal rights defined by scheme rules and the exercise of discretion by trustees or employers.
Both elements are hard to verify after the event. Member files may be missing or may require manual search. The position may be similar when it comes to accrual rates or retirement ages because rules often allowed for employer discretion and individual agreement.
The pragmatic answer is to treat these elements of benefits as data items. There’s a risk they are wrong. Pensioners are likely to challenge them at retirement or not at all. For deferreds, there may be errors but they are unlikely to be systemic. The risk can be mitigated by communications designed to flush out disagreement, whether by self-certification or at least benefit statements, and the risk can be covered by ongoing or run-off insurance . So this probably isn’t a major problem area and it’s not unreasonable for trustees to take the facts at face value.
The real problem is the legal rights that apply to the benefits after the point of leaving service. Ultimately what we need to know is the correct benefits payable. This is presumably the point of data cleansing.
Benefits are the product of facts and rights. Like with member information, work is needed if you want to confirm you know the legal rights. Unlike member information, it won’t be obvious if it’s missing or implausible – not unless you are a pension lawyer, and not necessarily even then.
Pension amounts are put in spreadsheets tranched according to the rate of pre-or post-retirement escalation that applies, or according to the payable age. These amounts look like facts. But they can’t be verified by asking members. They depend on an understanding of the legal entitlements. Scheme documentation can’t always be taken at face value.
Some elements look clear on the face of the rules but it is a mistake to rely on them. For example, normal retirement date may be irrelevant as it may well not determine – or not for everyone – the way that different tranches of benefit behave before 60 or between 60 and 65. Normal pension age used to be more important although it was always a term of uncertain legal meaning but in any case its relevance to deferred pensions changed significantly on 6 April 2005. Misunderstanding these terms can lead to paying incorrect benefits.
Benefit specification due diligence for bulk annuity transactions shows a consistent pattern. There are three main problem areas: the pre- and post-retirement escalation rates (“revaluation” and “indexation” respectively) and the payable age (the Barber window issue). In all three areas there is a statutory override which changed later. That was a recipe for disaster. The law about pre- and post-retirement escalation was always more complicated and nuanced than it seemed. It’s not just a matter of the index used and the percentage cap. And it has got more complicated since with the introduction of CPI. In fact the rule amendments filled in important points of detail and/or hard coded elements of statutory provisions that can’t now be removed.
In the meantime the administrators believed they were complying and trustees may not have taken legal advice on whether they were. If they did, the legal advisers may have made unconscious assumptions about admin practice rather than having detailed dialogue with the administrator. Both sides didn’t know what they didn’t know.
It is trustees who are responsible for this. Experience on buyout due diligence shows there are often ‘overs and unders’ and the net effect may well be a reduction in price. Problems can be stopped from getting worse and there are mitigations on– maybe underpayments of revaluation but overpayments in other areas such as late retirement uplifts after 60.
We are where we are. The decision point for trustees is whether they are going to do data cleansing once and do it well.
Read Anna Rogers’ article in Actuarial Post.
The views in this article are intended for general information purposes only and should not be used as a substitute for professional advice. Arc Pensions Law and the author(s) are not responsible for any direct or indirect result arising from any reliance placed on content, including any loss, and exclude liability to the full extent. Always seek appropriate legal advice from a suitably qualified lawyer before taking, or avoiding taking, any action. If you have any questions on the points raised in the above, please do not hesitate to get in touch.