Recent Posts



    News & Views

    Our latest news and blogs keep you in touch with our fast paced and ever changing industry.

    The Uniqueness Fallacy

    By Tom Nimmo on Tuesday, February 16, 2016

    Growing Pains

    Individuality seems to be something that nearly all tortured teenage souls strive for, but rarely obtain. Strangely enough, the same principle applies to pension scheme datasets, where unique identifiers for member records are frequently non-existent or just simply fail to be unique. For teenagers, the failure to fully realise their individuality might result in a heightened sense of ennui, but for pension schemes the consequences of identifier homogeneity can be far more dramatic.

    Without unique identifiers, matching data records consistently between tasks becomes time-consuming, with the increased risk that updates are applied to the wrong records or missed altogether. Therefore, the issues caused by unique identifier issues can have wide ranging and costly consequences for pension scheme administration practices.

    I got thinking about the issue of unique identifiers again recently due to the impending end of contracting-out in April this year. The event has created a real necessity for pension schemes to reconcile their contracted-out liabilities against HMRC’s records, and a key part of that process involves matching records. National Insurance Numbers (NINOs) form part of the identifying employee information within HMRC’s records, and they are frequently treated like unique identifiers in pension scheme administration datasets. Unfortunately, experience shows that NINOs are not entirely reliable when it comes to consistently matching records.

    The Uniqueness Fallacy

    The reason that teenagers struggle to be unique individuals and NINOs aren’t good unique identifiers can be explained by the ‘uniqueness fallacy’. The uniqueness fallacy is to argue that unique items within a group are unique in every respect. The reality is that teenagers and pension scheme administration records can be unique in some respects, but not others.

    When it comes to NINOs in pension scheme administration records, they simply aren’t unique identifiers and there are several reasons for this:

    • Pension scheme members can have multiple records within a dataset due to having separate periods of service in the scheme or holding benefits as both a full scheme member and a dependant beneficiary.
    • The member’s true NINO may be unknown, which can lead to dummy NINOs being created. These data items are inherently incorrect, potentially non-unique and unlikely to be permanent.
    • Systems that treat NINOs as both mandatory fields and unique identifiers can cause NINOs for members with multiple records to be changed intentionally, in order to work around the issues created by system constraints - see my previous blog for more information!
    • NINOs as they appear on pension scheme administration records are not particularly well controlled and so apart from some basic validation on the values, the NINOs on member records can just be plain wrong. Typos in the values are likely to go unnoticed until a reconciliation process flags an issue.

    From The Unexceptional to the Exceptional

    There is little that can be done about the issues that exist with NINOs, other than to follow best practice rules for both data management and pension scheme administration. As long as NINOs aren’t relied upon as unique identifiers there won’t be an issue. However, given that NINOs are one of the main personal identifying data items within HMRC’s records and many pension schemes are in the process of working on bulk contracted-out reconciliations, the importance of applying data management expertise to the process cannot be overemphasised.

    With matching records being just one of hurdles to overcome on the road to completing a contracted-out reconciliation, pension fund managers should ensure that they have accounted for the following process considerations:-

    • Data quality - Basic identifying information on pension scheme administration records needs to be accurate and complete.
    • Calculation routines - Benefit calculations and contracted-out amount comparisons need to be consistent.
    • Systematic checks - Analysis of the data needs to be conducted systematically and logically to identify and resolve issues as efficiently as possible.
    • Audit trail - A complete audit trail of every action and update to the datasets needs to be maintained.
    Getting it right is no small task, which is why there’s nothing juvenile about our contracted-out reconciliation service!

    Post has no comments.
    Post a Comment

    Captcha Image

    Back to News