Dark Pools Open Up About Surveillance Methods

Dark pool operators wary of bad behavior are retooling their software to protect the experience of institutional investors from suspicious trading activity.

“It’s part of the evolution of electronic trading,” said Adam Sussman, partner and director of research at Tabb Group. This gives dark pool operators “more granularity into the behavior and insight into participant behaviors and to prevent against other participants in the pool causing increased transaction costs for clients,” said Sussman.

But other sources said the moves are a response to concerns from buy-side traders who want transparency into dark pools which match buy and sell orders in the dark and whose participants are anonymous.

In March, Investment Technology Group announced the rollout of Liquidity Guard, a new breed of proprietary client safety measures that preserve execution quality and protect the ITG Posit dark pools.

“ITG has had a dark pool for a long time oriented toward institutional investors and quality was always a concern before toxicity became part of the lexicon,” said Jamie Selway, managing director and head of electronic brokerage at ITG. The firm felt it was time to make further investment and embarked on a year-long project, said Selway.

As an agency broker executing on behalf of client orders, ITG trades about 200 million shares a day, mainly through algorithms, which execute in the Posit dark pools as well as route out to non-ITG destinations.

Liquidity Guard sits on top of the ITG POSIT dark pools, which execute 100 million shares a day, of which 20 million shares are block trades that are executed in Posit Alert, according to Selway.

There are two pieces to ITG’s investment – one is an upgrade of existing quantitative techniques to prevent execution at bad prices, and the second piece is a more adaptive approach to detect the games that are being played.

Preventing executions at bad prices, Liquidity Guard compares current market activity against historical trading to identify abnormal pricing. “There are execution bounds that we won’t execute outside of,” explains Selway, adding, “Posit has guard rails around the prices against which you can execute.”

On the detection side, ITG has developed surveillance techniques for flagging suspicious activity that warrants further investigation by a dedicated analyst. “That further investigation may involve nothing, a call to the client to see what is happening or a filtering technique,” said Selway.

The new framework “is a systemization of a dedicated role towards analyzing these cases,” said Selway. “By doing that, we can ensure that we have real-time quality control that provides a feedback to our anti-gaming techniques,” said Selway.

Patrolling dark pools for unsavory trading behavior has always existed but broker-dealers are inventing new quantitative techniques for profiling client activity.

Last year Barclays LX rolled out its LX Liquidity Profiling Framework, a sophisticated surveillance system to monitor trading activity in its dark pool, which became the second largest dark pool in January.

“Traditionally dark pools and ATSs have offered some degree of classification to participants within the pools,” explained David Johnsen, a director in Barclay’s Equities electronic trading product. Institutions had the option of interacting with all of the liquidity, just the broker dealers, just the institutions, just the high frequency traders. “You could opt in or opt out based on this classification,” said Johnsen. But the problem with painting a broad brush of client type is that within that category not necessarily all the participants behave the same, noted Johnsen. “Institutions could be the most toxic order flow, depending on the institution," he said.

Among the first to take this approach, Barclays came up with an empirical measurement of how clients actually trade, which takes into account their overall performance, their short-term alpha and the overall aggressiveness of their order flow. "With this empirical methodology, we can give everyone a score from the most passive, least toxic client, all the way up to the most aggressive, short-term alpha client,” said Johnsen. “With Liquidity Profiling, the client can option into as much of the pool as they want to, order by order, based on the quality of the order flow they want to interact with,” he said. Essentially, there are five different buckets to choose from and the client can opt to interact with all five, or choose the most passive 1,2,3, for instance.

So what’s behind the latest investments in surveillance of dark pools?

“When we looked at the state of the industry there is fear of gaming and question being raised about some other dark pools and headline risk of what dark pools are doing,” said Selway.

Institutional investors that anonymously execute block trades in dark pools are asking more questions of the brokers after a series of regulatory actions.

In 2011, the SEC fined Pipeline Trading System $1 million for allowing a trading affiliate to match against institutional orders, unbeknownst to them. Then in 2011, LevelATS, whose owners are many Wall Street firms, settled charges over sharing information with Lava Trading, a technology affiliate of Citigroup.

1 of 2