Financial exchange regulators use computers to crack traders' code
Regulators trying to translate exchange dealers' clever chat into plain English as they probe whether they are colluding to rig key rates
"What's your interest in Bill and Ben in the pick?"
While most people wouldn't understand the question, this is common vernacular for London currency traders active in the daily 4 o'clock "fixing" of global reference exchange rates.
A basic translation of that particular line of Cockney rhyming slang, originating in London's working class East End, would run something like: "Are you a buyer or a seller of dollar/yen at the daily benchmark rate-setting fix?"
The wisecracking, rhyme-based language that can make a "whistle" mean a suit (from "whistle and flute") has been used by traders in the City of London for decades. But can a computer crack the code?
Financial regulators are now trying to translate such exchanges into plain English as they investigate whether dealers have colluded to rig benchmark foreign exchange rates used to price trillions of dollars' worth of deals.
Britain's financial watchdog, the Financial Conduct Authority (FCA), has about 60 staff dedicated to the foreign exchange investigation, some of whom are former market participants who are helping to decode the traderspeak.
Banks now want to go one step further, and are looking at acquiring Big Brother technology that can spot and prevent inappropriate communication or fraudulent activity rather than piecing together evidence afterwards.
"There is a huge market for this right now," said Sang Lee, founding partner at Aite Group, an independent research and advisory firm focused on business, technology and regulatory issues in Boston.
Lee said the focus was on software that could monitor and understand the language traders spoke in their electronic chats, instant messaging and e-mails, and alert the bank to potentially inappropriate communications or nefarious activity.
"For at least 12 months now, this has been one of the major focus areas for both market participants and the technology companies," he said.
Traders at banks and other financial institutions often communicate with each other via third-party chat rooms.
The world's biggest banks, Britain's FCA, the Bank of England and the US Department of Justice are among those trawling through countless chat room transcripts for evidence of questionable activity as part of a global investigation into the market.
Online communications featured prominently in a separate five-year inquiry into manipulation of an interest rate known as the London interbank offered rate, or Libor.
One of the technology companies that has gained from banks' increased vigilance is Digital Reasoning, based in Nashville, in the US state of Tennessee. Digital Reasoning's software can scan up to billions of communications from thousands of traders, spotting language patterns and raising a red flag to potentially dubious activity.
Rob Metcalf, the company's president and chief operating officer, said there had been a sharp increase in interest for this software from banks over the last six months, just as the global investigation had taken off.
"The big banks will drive this, and regulators are interested too. It will be a baseline capability across major banks in two years," Metcalf said.
He said his firm was talking to 10 to 20 of the world's leading financial institutions, particularly American and European firms.
Banks had been moving towards more sophisticated analytics across all forms of electronic communications over the past year, said Michael O'Brien, head of sales for the SMARTS Broker market surveillance platform, part of Nasdaq OMX.
One area they are looking at is "relationship forensics", tracking traders' communications with each other to spot patterns that point to a close working relationship. These could include behaviour such as always copying a particular person on e-mails, or if someone always responds to an e-mail within minutes. O'Brien said that was because any dubious activity between two or more people was bound to be carried out by those who had a close relationship with each other.
Banks are looking for software that is able not only to detect language patterns in traders' electronic communications, but to match that up with patterns detected across other media, such as social media profiles, Aite Group's Lee said.