Market Police Deploy New Algorithm Weapons Against Spoofers
(Bloomberg) -- The day after the Federal Reserve raised interest rates, volume was so heavy on the Nordic stock exchanges that Joakim Strid’s team of analysts identified 500 instances of possibly fraudulent trading -- and that was before lunch.
“It’s been a very busy morning,” said Strid, the head of European surveillance at Nasdaq Inc.
Strid’s Stockholm-based staff was keeping an eye on trading with Smarts, a software program Nasdaq uses to unearth suspicious behavior on 13 exchanges in Europe and the tech- driven exchanges it manages in the U.S. Alerts pop up on terminal screens that indicate possible “spoofing,” an illegal scheme in which traders drive the prices of stocks up and down with bogus buy and sell orders.
“The next step is to conclude whether it’s worthy of further investigation,” Strid said.
Nasdaq’s market technology business produced $59 million in net revenue in the third quarter, and Smarts has emerged as the No. 1 trading-surveillance player in the world at a pivotal moment for the burgeoning industry. Allegations that Navinder Singh Sarao, a 37-year-old day trader near London, contributed to the flash crash in May 2010 by spoofing the futures market is just one of a series of cases that have focused the investing world this year on a specific breed of algorithmic fraud.
With regulators pressing market participants to root out securities scams, Nasdaq and its rivals are jockeying to install the equivalent of closed-circuit TV cameras on trading platforms around the world to monitor and record every gyration that a security or a derivatives contract makes.
Cinnober Financial Technology AB, also based in Stockholm, provides market surveillance to exchanges ranging from the Frankfurt-based Deutsche Boerse to the Dubai Gold and Commodities Exchange. Vertex Analytics helps the CME Group Inc. and the Chicago Board of Trade scrutinize the futures markets. And London Stock Exchange Group Plc uses its own Millennium program to police its U.K. exchange. In 2016, the $450 million trading-surveillance market may increase as much as 10 percent, according to Aite Group, a Boston-based consulting firm.
“Every time there’s a high-profile case, there’s more spending on surveillance systems,” said Danielle Tierney, a senior analyst at Aite.
While regulators have long relied on whistle-blowers and spreadsheets to detect abnormal trading behavior, these methods can’t handle the firehose of data that now floods the markets. There were as many as 300 million price-quote changes on a single trading day in the U.S. stock markets in 2015, an eleven- fold jump from 2005, according to Credit Suisse Trading Strategy data. It’s easy for fraudsters to slip phony bid and ask orders into the flow, said Michael O’Brien, head of product development at Smarts.
“We have to capture every trade now,” O’Brien said. “In today’s markets it’s all about analyzing patterns and contexts.”
Yet given how rapidly fraudsters can change their methods to hoodwink human beings, outwitting surveillance software could be even easier. Algorithms are sophisticated but they’re incapable of determining whether a flurry of buy and sell orders are legitimate or unlawful.
“The surveillance tools are merely the first line of defense,” said Haim Bodek, founder of Decimus Capital Markets, a New York-based algorithmic investing firm. “These tools can help bring suspicious activity to the attention of regulators, trading venues and brokers, but they’re a poor substitute for a compliance program that monitors activity across affiliated accounts and groups of traders.”
Even when the technology does uncover possible skulduggery, it takes years to adjudicate cases because prosecutors need to prove the intent of traders. In December 2010, surveillance analysts in the London office of Bats Global Markets Inc. used Smarts to figure out how a trio of traders in Hungary were spoofing the shares of a small Australian mining firm. The spoofers drove down the stock price with phantom sell orders, scooped up the shares, and then goosed the price with phony bids, according to a civil suit brought against them by U.K. regulators.
“They were trying to suck people into the marketplace,” said Michael Aitken, a professor of information and communications technology at Macquarie University in Sydney who founded Smarts in 1994.
It took five years to wrap up the case. On Aug. 12, a British judge ruled the traders had violated market-abuse laws and fined them 6.1 million pounds ($9.1 million).
However long these cases take, the cat-and-mouse game between watchdogs and spoofers is poised to intensify next year. Cinnober and Scila AB, another Stockholm-based player, are installing a trading-surveillance system at the New York Stock Exchange that will replace many of the monitoring functions performed by the Financial Industry Regulatory Authority.
Aiken said it’s only a matter of time before spoofers begin manipulating the prices of securities and derivatives contracts in multiple markets at the same time, including dark pools.
“If you can influence prices in one market and trade somewhere else, then let’s see who catches you,” Aitken said.