Illegal money in the international banking system has become a pressing issue for governments and the world’s financial institutions.
To take just a couple of recent examples, there’s the money-laundering scandals that have engulfed some of the biggest banks in the Nordics, while recent figures from the Home Office estimate that as much as £90 billion is laundered through the UK each year.
No wonder HMRC and the banks are stepping up their efforts to combat these illegal activities. The fallout has also led to tightened regulatory requirements in many areas, as well as the threat of very large fines for firms that are viewed as not having done enough to combat the issue.
Anti-money laundering is, however, a major task for firms. Why? Because in order to spot illicit dealings, firms need to know where the funds are coming from and where they are going to. But instead of moving money directly from point A to point B, money-launderers divert funds to a dozen other points in between. Notoriously, of course, they do this in smaller sums in order to bypass the regulatory limits on the amount of money a customer can send or receive — a technique known as ‘smurfing’.
Understanding the maturing role of graph databases in the enterprise
Unfortunately, many traditional anti-money laundering technologies aren’t designed to connect the dots across the many intermediate steps. That means that detecting money-laundering requires a tremendous amount of manual effort, sifting suspicious transactions out from legitimate ones in the huge volumes of data that flow through banking systems every day.
Teams of inspectors can spend months going through reams of such material, and too often, a bank’s standard instruments for dealing with the problem – such as monitoring for deviation from normal banking patterns — is all about discrete data, which can’t easily detect the shared characteristics that typify money-laundering networks. Furthermore, many discrete methods are prone to false positives, possibly as many as 95%, which can negatively impact customer relationships and result in lost revenue.
It’s the needle in a haystack issue, for sure. But while no money laundering measures can be comprehensively effective, the good news is that opportunities for improvement can be achieved. We can do this by looking beyond the individual data points to the connections that link companies, accounts and transactions.
Gartner: top 10 data and analytics technology trends for 2019
Misconception: relational databases cannot actually identify links in data
A business database technology called graph software has the power to uncover these connections and is being used already to help detect and prevent money-laundering. It’s the same graph technology that helped data journalists at the International Consortium of Investigative Journalists, the group behind the Panama and Paradise Papers, to analyse and structure leaked data about business entities, wealthy individuals and public officials. We would not have discovered how the elite avoid and evade tax if it wasn’t for their work cracking hidden money trails concealed in data using graph data-based techniques.
Graph technology works so well because the standard way of working with information in business – relational databases — were not created to identify links in relationships. Such queries are technically tricky to build in a relational database and expensive to run in complex money laundering scenarios. Making them work in synchronous time is problematic, with performance eroding very fast as the total dataset size increases.
But a graph database is practically impossible to beat at mining connections in complex datasets and analysing the relationships between a large number of data points. At the same time, graphs are supremely fast at exposing patterns, giving financial institutions real-time insight into their assets and associated relationships. For example, in 2016, Dun & Bradstreet (D&B), the world-leading technology and data group, adopted graph technology to better support a new B2B compliance service for customers that want to investigate all historic company ownership records linked to individuals, in line with new international transparency regulations designed to counter tax evasion and money laundering.
For the company’s senior compliance manager, Paul Westcott, “The new ‘beneficial ownership’ mandate calls for highly trained staff, and this activity is hard to scale. A single query might tie up key people for 10-15 days, resulting in lost revenue.”
By using graph software’s ability to efficiently map shareholders’ interests globally – despite complex inter-company dependencies — he and his team help clients accelerate compliance and book new revenues without delay: potential hours of manual research by expensive professionals has been avoided, and D&B’s associated business activity is expanding at a double-digit rate.
“Being able to quickly understand relationships between data gives us the ability to rapidly interpret corporate structures and any dilution of ownership of a business,” Westcott explains. “[Graph tech’s] network of nodes and connections mean data can be surfaced for an individual in milliseconds – a very quick return of information, and an ideal fit for our needs.”
It is imperative that chief security officers in any organisation, with a responsibility to head off money-laundering, have a tool that can help them fight money laundering effectively — and graph technology is emerging as one they can rely on.
