The consensus when investigating financial crime alerts for individuals, is to close all alerts where the date of birth is more than 2 years apart. A more aggressive strategy is to close alerts where the dates of birth are different – even if only by 1 day. In this short article, we will address the merits of these strategies.
First, the gain:
Eliminating hits outside of a 2-year span on either side will typically eliminate 90% of alerts where dates of birth exist
Eliminating where the dates of birth are not the same will reduce your workload by 99.99%.
Undoubtedly, there will be multiple alerts where a date of birth is not present or only a year of birth is present. So, the gain can be significant, but what might be missed? Surely if the dates of birth are different, then it is a different person? You are in control of the dates of birth captured in your KYC process, but you are not in control of the list data, so what are the possibilities of the list dates being inaccurate?
1st of the month dates – as can be seen in the graph below, list dates of birth skew towards the first of the month, and skew even more towards the first of the year. This is because list dates are often captured on a year and month only basis and can be recorded as being on the 1st of the month rather than having no day. This can impinge on your investigation.
Day month reversals – when the day of the month is the 12th or less, day and month reversals can occur, this can often happen where American and European format dates are interchanged.
Multiple dates of birth – in some cases list entries are provided with multiple dates of birth, this demonstrates that a definitive date is not known, so placing 100% reliance on the given dates of birth is a leap of faith.
Incorrect dates – sometimes the listed date is just wrong, we have seen many occurrences of dates of birth being wrong in what turns out to be a true match. However, it is important to bear in mind that where a list contains a date of birth that is more fiction than fact, eliminating hits outside the span of two years alone does not necessarily protect you, it could easily be 10 years out, or 50 (decade year reversal is a possibility).
It is important to note that other factors apply when a year of birth is present rather than a full date of birth, and there are other considerations to be mindful of with regards to dates, but these are the ones that affect the argument of which is better, eliminating hits outside of the 2-year range vs an exact approach.
In conclusion, a 2-year date of birth range can give you some added protection, but not full protection. An exact date of birth approach must be tempered with the knowledge that there exist known quality issues that must also be addressed – such as the first of month date. We would never recommend using an exact date of birth approach without addressing the issues listed above.
The decision is all about risk and reward. The different types of screening alerts – sanctions, PEPs, adverse media – all have different risks and may need different approaches.
