Jerome Kerviel, a former trader at Societe Generale in Paris whose rogue trades cost $6 billion to unwind in 2008, faces five years in prison and a fine of nearly half a million if convicted by the three-judge panel that heard his case early in the summer. The trial laid bare countless ways he could or should have been caught, despite his assertions that his superiors knew what he was doing. But as Kerviel awaits news of his fate, there's much talk of what banks can do to avoid SocGen's errors when it comes to risk management.

One way to stop a rogue traders in their tracks: improving trading and operational risk controls which even SocGen's internal investigation now concedes were sorely lacking.

But there is yet another even more novel approach now being proposed by Cataphora, an eight-year-old Redwood City, Calif., software firm specializing in evidence analytics. The approach involves monitoring unexplained changes in an individual's behavior or from the controlled norm- meaning the activities of his immediate colleagues-by analyzing the emails, other digital communications and electronic records for their content as well as their broader context.

This includes studying the content and context of the messages, when they were sent, how many were sent and to whom, and how often they were sent. The software can also evaluate whether a person's typical daily workflow differs from that of his colleagues with the same responsibilities.

The analysis results in more than just cold facts. Designed and built by Cataphora's mathematicians, linguists and software engineers, the software produces alerts and data visualizations that can be interpreted by a firm to determine if an employee's conduct needs to be investigated. The potential users of such software: legal and compliance departments as well as operational risk management and IT.

"Traditional compliance and risk management systems rely on fixed, predetermined risks which behave in a predictable fashion," says Elizabeth Charnock, Cataphora's CEO. "That means they depend on the rule creators' ability to anticipate the forms wrongdoing might take (in order) to prevent it."

Cataphora's software also can find close social ties between employees and to look at a company's so-called shadow social networks. While most shadow social networks aren't anything to worry about, they can sometimes be instrumental in helping employees perpetrate fraud. Shadow social networks can be especially useful to wrongdoers in companies that have stringent fraud-prevention mechanisms, such as requiring multiple signatures for large transactions. In those cases, it often takes groups of people to skirt controls, Charnock says.

While Cataphora's software in its current form could not have detected Kerviel's wrongdoing, the company plans to make advance detection possible in the future. The firm is designing separate software based on the same technology that can provide a real-time analysis of electronic messaging and workflows that could be used by either compliance or operations departments.

Here are red flags that real-time behavioral modeling would have picked up from Kerviel's communications: Overriding chain of command. While Kerviel's decision to exceed trading limits alone might not have raised a red flag if colleagues were also doing the same thing, the fact that he violated the hierarchical chain of command was unusual. Excessive and inappropriate communications with the same broker: Kerviel relied almost exclusively on the same broker, Moussa Bakir, a former futures broker at SocGen's Fimat brokerage unit, to place orders. There are a total of 2,000 pages of instant messages exchanged through Reuters Messaging. "It would have been normal to execute orders through whoever picked up the phone to take the order rather than using the same broker repeatedly," says Charnock.

Consistent language for making confirmations. When Kerviel allegedly created bogus confirmations of transactions he repeated the exact wording across numerous electronic communications.

Consistent approach to recording fake trades and cancelling positions. All of the phony transactions Kerviel generated had the same three characteristics-a significant offset, meaning a value date later than the date of the transaction, the use of internal counterparties within Societe Generale or a small external counterparty. In all cases, the trades were cancelled before their date chosen for their valuation.

Change in language on activities. The ultimate clincher. Even if Kerviel's supervisors, compliance executives or the operations department didn't catch on to Kerviel's wrongdoing, Kerviel himself expressed concern about his own conduct and his communications. Here is just a tidbit of the exchange between Kerviel and Bakir. On Oct, 11, 2008, Kerviel wrote: "You haven't said anything about our trades? If so, I'll smash your face in." Response from Bakir: "It's between you and me." On Dec. 13, 2008: Bakir wrote "You are absolutely going to have to take your holidays." Kerviel replied: "Behind bars." Had SocGen unraveled his trades earlier it might have mitigated its losses. When SocGen began unwinding Kerviel's positions on Jan. 21, the U.S. markets were closed which put pressure on futures markets and exacerbated its losses.

This article can also be found at

Register or login for access to this item and much more

All Information Management content is archived after seven days.

Community members receive:
  • All recent and archived articles
  • Conference offers and updates
  • A full menu of enewsletter options
  • Web seminars, white papers, ebooks

Don't have an account? Register for Free Unlimited Access