Banks are hunting for rogue traders the same way the Pentagon hunts for terrorists: By predicting their behavior.
Instead of looking for weapons deals and on-the-ground movement plans, predictive algorithms can spot shady bankers by tracking their emails when they talk about tickets to sports events and concerts. Cutting-edge analytics systems frown upon choice seats to the Super Bowl, which raise red flags for bribery.
“By understanding tickets, events and related content in the context of email conversations with government officials, the system can identify language that may be indicative of bribery… or at least that require further human review,” said Rob Metcalf, chief operating officer of Digital Reasoning.
Digital Reasoning started off doing large-scale analyses of human communication for the U.S. government, in particular terrorists operating around the world. That work ended up being a perfect fit for banks that had already been keeping a close watch on their traders’ accounts but not their interactions.
Metcalf said the analysis of linguistic patterns used by bankers and terrorists are key indicators that can be used to identify possible problems. Teaching the system to understand and identify value — such as tickets to events — can help trigger the alarm bells.
On Wall Street, where investors frequently hear that “past performance is no guarantee of future returns,” a truth has been revealed: Algorithms that track people are far more predictable than algorithms that track stocks. Algorithms that track people are far more predictable than algorithms that track stocks. Software, made by companies including Digital Reasoning, can scan employees’ phone call transcripts, email and chat to form a basic understanding of human behavior— enough to predict if they might break the law or make a series of risky bets.
Think IBM’s Watson, but as predictor of shady traders.
“What large financial institutions in particular were seeing was that the human risk areas they were getting penalized for were ultimately coming back to people’s communications,” Metcalf said.
Spotting rogue traders shouldn’t be hard, even for human beings, because Wall Street workers can be cartoonishly villainous in their digital communications. One trader who was arrested this week for allegedly causing the 2010 flash crash suggested in an email that he traded “by instinct” rather than reason. Another trader who rigged the LIBOR interest rate set himself weekly Microsoft Outlook reminders to make sure to rig the numbers every Monday.
Similarly, Fabrice Tourre, a Goldman Sachs trader who was investigated for his role in the subprime mortgage crisis, embedded incriminating critiques of Wall Street in his love letters to his girlfriend.
An $825,000 fine and a fraud conviction that he did not appeal was the result of Tourre’s tormented emails, which included deep thoughts like: “When I think that I had some input into the creation of this product (which by the way is a product of pure intellectual masturbation, the type of thing which you invent telling yourself: “Well, what if we created a “thing”, which has no purpose, which is absolutely conceptual and highly theoretical and which nobody knows how to price?)”
Read the full article here.