This content is from: Home

Money Laundering - Tech detectives

The world's banking regulators are stepping up their war on money laundering. Unfortunately, it's an attack that will cost even law-abiding financial institutions time and money.

The world's banking regulators are stepping up their war on money laundering. Unfortunately, it's an attack that will cost even law-abiding financial institutions time and money.

By John Wagley
August 2001
Institutional Investor Magazine

Banks and brokerages have been enlisted as the first line of defense - a high-tech detection system -in identifying suspicious transactions, particularly when handling foreign currencies and dealing with wealthy clients. Under the U.S. Bank Secrecy Act and similar laws in other countries, financial companies must file reports on large cash movements and call authorities' attention to questionable transfers. And the rules have teeth, as U.S. Trust Corp. of New York learned last month when it was hit with $10 million in fines for compliance lapses.

The task has grown so complicated and expensive that financial firms are increasingly relying on technology, both to file transaction reports and to finger the crooks. Indeed, no one seems to know the full costs of policing and deterrence. In recent congressional testimony, U.S. Treasury Secretary Paul O'Neill admitted that his department hadn't figured out how much the U.S. government spends on money laundering investigations. A big bank spends $15 million a year to meet antilaundering requirements, says Charles Intriago, a Miami-based lawyer and publisher of the "Money Laundering Alert" newsletter. He says that the U.S. banking industry shells out hundreds of millions of dollars a year on compliance and countermeasures.

Money launderers have grown adept at turning ill-gotten gains - whether from drug sales or arms trafficking or political corruption - into cash alternatives such as money orders or traveler's checks. With their money in cash equivalents, they can go to another country and wire it to still other locations, usually in a variety of currencies. By the time the sums are ultimately withdrawn, often as loans or electronic funds transfers from a legitimate bank, their origins are virtually impossible to trace.

The International Monetary Fund estimates that $590 billion flows each year through banking channels in ways that cover up criminal or terrorist activities.

The new defenses take the form of databases and analytical software, using advances in artificial intelligence and pattern recognition to screen customers' financial histories and scan their transaction records. It's all aimed at sniffing out the few illicit transactions from among the millions of legitimate ones before all hope of traceability is lost.

Deutsche Bank, for example, has gone so far as to develop its own high-powered screening technology. Thomas Obermaier, the head of risk management for the bank's global cash management business, says that he considered buying such systems from software houses, but Deutsche deemed the available products "too general" for its multinational requirements. "On the whole," Obermaier says, "vendors tend to lack banking and securities experience."

Vendors, of course, beg to differ. London-based Searchspace, which supplies risk management and compliance software to Bank of New York and Royal Bank of Scotland, among others, says that its laundering detection system is combing through 30 million transactions a day for one recently signed client, which has not consented to have its name disclosed. Konrad Feldman, chief executive officer of Searchspace's U.S. subsidiary, says that his company is selling to "top-tier financial services organizations" for use in retail, private and correspondent banking departments.

Whether homegrown or off-the-shelf, the technology is light-years beyond where it was in 1970, when the Bank Secrecy Act first required paper reports on cash transactions of $10,000 or more. In 1990, with money-related crimes on the rise and those reports too numerous to fully examine, the U.S. Treasury created the Financial Crimes Enforcement Network. The idea was to use technology to sort through the documentation and counteract the increasingly sophisticated tactics of the money launderers.

FinCEN and its counterparts around the world routinely share information and technologies. The multinational initiatives have culminated in threats by the Financial Action Task Force on Money Laundering - a 29-country organization formed 12 years ago by the Group of Seven industrialized countries - to impose economic sanctions on nations that don't clean up their banking practices. Egypt, Nigeria, the Philippines and Russia are among those now on the brink of pariah status.

Financial institutions have two basic defenses against money laundering: assessing the risk of doing business with a client at the time an account is opened and monitoring and tracking existing clients' transactions.

Deutsche Bank has designed a system to create a risk score when it is vetting the credentials of a customer opening a new account. Akin to the credit scores that help bankers decide whether to make a loan, risk scores are based on past business activities and other probability factors. But to be effective, the score has to be derived from good data.

Because the quality of commercial databases varies from country to country, Deutsche Bank prefers to rely heavily on information gathered through its 107-country banking network. "The information is reality-tested, because it is based on our own risk experiences," says Obermaier, a former attorney in the U.S. Justice Department's criminal division who joined Deutsche Bank in New York in 1996. "We also own the quality control since it includes information that we have tested. We know the source, have ascertained it is reliable and believe it to be of a certain quality."

The bank went live last month with a prototype of its system for the account-opening stage. The most critical challenge, Obermaier says, is to prove that the theory works in practice - that the mathematical algorithms yield accurate risk probabilities.

Deutsche Bank previously developed adaptive software, called db Tracker, that follows customers after banking relationships are in place. The system "learns" from each transaction experience and automatically adjusts the risk score. If that score passes a critical point, it triggers an investigation.

Since implementing db Tracker last year, Deutsche Bank has reduced the number of full-time employees monitoring for money laundering from 50 to four, says Obermaier. He declines to discuss whether the technology has rooted out criminals but says that it has detected some "unusual activity."

Customer tracking of this sort is two to three years old. The first systems were static, merely flagging deposits above a certain amount. "The trouble with those was that launderers tend to be pretty smart," says Searchspace's Feldman. "They can size their transactions accordingly." Technologists responded by programming in more "expertise" to recognize the skulduggery.

ACI Worldwide, a unit of Transaction Systems Architects in Omaha, Nebraska, offers a tracking product called Prism that uses a type of artificial intelligence known as a neural network. Prism takes a comprehensive picture of transaction activity across multiple business units, which should appeal to large, diversified organizations, says ACI product development chief Brian Fisher. "One thing we're hearing from regulators is that banks should look at the complete customer relationship. Whether it's a correspondent account, private banking or the brokerage area, we take feeds from each payment mechanism," he adds.

Prism zeroes in on what Fisher calls the first two stages of money laundering: placement, when money is deposited into an institution; and layering, when the money is circulated through multiple accounts. By the third and final stage, when the funds are fully integrated into and careering through the banking system, the perpetrators are beyond the reach of computerized intelligence.

Technology experts stress that scoring and tracking methods must be continually improved to stay ahead of the criminals. "Money laundering strategies tend to change along with the technology meant to combat it," says Daniel Brooker, executive vice president at the National Fraud Center, a Horsham, Pennsylvania-based unit of Lexis-Nexis Group that does background checks on bank clients. "The baddies find a system's cracks, and we need to adjust."

The bad news is that there are more baddies than ever - and plenty succeed in cleaning their money. Richard Rahn, a former chief economist of the U.S. Chamber of Commerce, argues that antilaundering laws have had no noticeable impact on drug running, tax evasion, terrorism or other stated targets. Including bank compliance expenses, a money laundering conviction - of which there are fewer than 1,000 a year - costs the U.S. private and public sectors at least $10 million on average, Rahn estimates.

A self-described libertarian who heads Novecon Financial, a Washington-based consulting firm, Rahn recommends abolishing the laundering statutes; he contends that other laws can accomplish the desired ends at less cost. But he adds, "There is a need for the financial industry to devise systems to prevent fraud." They could be the launderers' worst enemy.