Cutting the terrorists' money supply: The role of IT

When President George W Bush froze the assets of 27 organisations and people linked with Osama bin Laden the problem of money laundering came to international attention. But for banks and law enforcement agencies it's a problem they've been fighting for years. Sally Watson takes a look at how technology can hinder and help them.

Estimates on the amount of money laundering going on across the world vary wildly. Most experts put it somewhere between $1tr and $2tr annually, but openly admit no one really knows how much passes through the system.

It is hardly a new problem. The process of making dirty money clean has been going on for years but the advent of sophisticated technology designed to make global commerce quicker and more convenient for customers has also aided the money launderers themselves.

TowerGroup Research recently estimated $500bn of laundered money passes through the US every year. What's more, the problem is getting worse.

"Technology and the interconnectedness of global banking organisations have increased the international effects of money laundering," said Breffni McGuire, analyst at TowerGroup.

"In particular, emerging internet technologies and evolving electronic commerce banking services are likely to accelerate development of new types of laundering activities."

According to the National Criminal Intelligence Service there was a 27 per cent rise in the number of suspicious financial transactions in the UK last year. Over 18,000 cases were reported.

But head of its economic crime unit, Andy Blezzard, said out of 575 registered banks in the UK only 170 - less than 30 per cent - made any suspicious transaction reports last year.

"Financial institutions have a legal duty to look for criminal activity," Ruud Nijs, European MD at HNC Software points out. "Most banks still have manual systems where large amounts of cash are flagged up by cashiers and then checked."

This standard system is used worldwide. In the UK it is up to individual banks to decide what to check, but in the US any cash transaction over $10,000 comes under suspicion. In the Netherlands, it's NLG25,000 (approximately £7,000). These thresholds are well publicised and anyone wanting to hide financial activities can easily divide cash amounts so that they stay below the limit.

The problem facing the financial industry is huge. To follow a series of relatively small cash amounts through a global system which deals in trillions of dollars per day, via possibly hundreds of different routes, is somewhat akin to finding a needle in the proverbial haystack.

"We are talking about vast volumes of data, millions of customer records," said Peter Dorrington, business solutions manager at software company SAS Institute. "You can bounce records around very quickly, and organised crime knows how to do this very well."

The simple method is to freeze suspicious accounts to break the chain of individuals and institutions filtering the money, the same method used by President Bush in his attempts to squeeze the Al Qaida network.

Nick Downes, banking consultant at Logica, says the principle of so-called criminal 'hotlists' is now well accepted, as long as law enforcement agencies can come up with enough evidence. "Two or three years ago the OFAC [Office of Foreign Asset Control] regulations were only really used in the US. Now the global system is being tightened up, and the US government is saying banks won't be able to do business in the US unless they sign up," he said.

But it's a reactive system that relies on intelligence which is notoriously difficult to gather, and increasingly sophisticated data mining techniques are now being used to identify patterns of criminal behaviour proactively.

Both HNC and SAS have developed products based on previously successful fraud and CRM systems. "The technology is very similar to techniques used in marketing," admitted Dorrington. "You're trying to identify behavioural and transactional patterns."

Data profiling helps build up a picture of normal and abnormal financial activity. It might establish, for example, that high-worth customers from one geographic region tend to make a certain number of transactions per month - a sudden increase would then become suspicious.

Where data from proven cases of money laundering or terrorist activity can be used, the software can build a predictive model which can be run against customer and transactional databases in order to identify matching activity.

And it's not just useful for the financial industry. According to HNC's Nijs, data mining technology provides basic components which can be used anywhere. "We're talking to airlines in North America and the insurance industry. As long as we have an availability of good data we can help solve these problems," he said.

The tightening of regulation in the US and Europe has driven much of the money launderer's business offshore. Banks in South East Asia and the Middle East now offer the same 'no questions asked' policy once famously offered by Swiss institutions.

These banking black holes are a problem for investigators, but not impossible to solve. "You can pay much more attention to transactions going in and out of secretive banks than a normal household bank," said SAS' Dorrington. "It draws attention to itself."

Evidence of the success of these systems is difficult to measure. Few banks are willing to reveal the methods they use or the results they achieve. Nijs estimates that for every 20 cases HNC's software identifies as suspicious, only one or two will turn out to be genuine criminal activity.

Dorrington claims the successes are there but they can be difficult to quantify. "We're looking at a spider's web of activity and identifying the weaknesses so we can snip out those links," he said.

"We may not be able to detect the terrorist at the end of the chain or his funding source at the beginning, but because it's a chain we can take out a single link and it fails."

Whatever the results, from the end of November new FSA rules will require anyone selling financial services in the UK to take the same precautions as banks.

Online finance companies, independent financial advisors and building societies will have to appoint a money laundering officer to monitor and report suspicious activity. They'll also have to prove they've taken appropriate measures - including statistical analysis - to carry out investigations.

Ultimately, most governments and law enforcement agencies will admit the problem is impossible to eradicate, with or without data mining techniques. But with an increase in the use of automated payment systems, digital cash and ecommerce, technology will have to be increasingly used to solve the problem it creates.