In the latest round of regulatory action involving high frequency trading and dark pools, the SEC announced yesterday that it reached a settlement with ITG, Inc., and its affiliate Alternet Securities, Inc., imposing a $20.3 million sanction based on ITG’s misuse of confidential order information to benefit the firm’s proprietary high-frequency trading.

The SEC’s Order states that ITG historically marketed itself as an “agency-only” brokerage firm that did not engage in proprietary trading for its own account. ITG customers included “buy-side” customers, such as investment advisors, asset managers, pension funds and hedge funds, and “sell-side” customers, principally broker-dealers. In addition, ITG operated POSIT, an alternative trading system (often referred to as a “dark pool”) that anonymously matched and executed orders received from ITG customers and other POSIT subscribers.

As part of its brokerage business, ITG provided services to its customers through a suite of trading algorithms and smart order routers. These algorithms and routers broke up large customer orders into smaller “child” orders and sent the orders to an execution venue determined to be most appropriate — either POSIT, another other dark pool, or an exchange.

In early 2010, the firm explored the possibility of expanding its business and established a limited proprietary high frequency trading operation, called Project Omega, as an experiment to determine whether the business could be profitable. At that time, ITG issued policies, reflected in multiple memoranda from the compliance department, prohibiting the proprietary trading operation from accessing information about ITG customer and POSIT orders.

Contrary to these compliance directives – and, according to the SEC order, without the knowledge of senior ITG management – the ITG personnel responsible for implementing Project Omega did, in fact, obtain access to data feeds containing confidential order information from POSIT as well as information regarding ITG customer executions in other trading venues. This data was then used to drive two trading strategies implemented by the Project Omega traders.

First, a high frequency algorithmic strategy called the “Facilitation Strategy” used the POSIT order data to identify instances where a sell-side POSIT subscriber was willing to trade at a price less favorable than a price currently available in the market. In those instances, Project Omega opened positions in the market on the same side as the POSIT subscriber, and immediately thereafter traded against the subscriber in POSIT at the less favorable price. For example, if order information in POSIT indicated that a subscriber would be willing to buy a particular security at $10.02 per share, and Project Omega could obtain that security elsewhere in the market at $10.00 per share, Project Omega would buy the security at $10.00 and then sell it to the subscriber through POSIT at $10.02.

Project Omega also employed a second high frequency strategy called the “Heatmap Strategy,” which used data about executions ITG customers had obtained in dark pools other than POSIT.  Based on that information, Project Omega determined which dark pools had provided executions to ITG customers at the midpoint between the best bid and offer suggesting that additional liquidity existed in that dark pool at the midpoint price.  In those instances, Project Omega would open positions at the best bid or offer in a displayed market (such as NASDAQ, the NYSE or another exchange), and immediately thereafter close the positions in the identified dark pool at the more favorable price.  For example, if the Heatmap detected midpoint executions (at $10.01) by an ITG customer selling securities in an external dark pool when the best bid was $10.00 per share and the best offer was $10.02 per share, Project Omega would buy the security at $10.00 per share in a displayed market and then sell the security at the midpoint ($10.01) in that external dark pool.

The SEC charged ITG with fraud under Section 17(a)(3) of the Securities Act based upon its failure to disclose the nature of its proprietary trading activities. The SEC also charged the firm with violations of Regulation ATS based upon its failure to establish adequate safeguards and oversight procedures to protect the confidential trading information of POSIT subscribers and to amend its Form ATS to reflect its proprietary trading.

The SEC order noted that the team responsible for developing Project Omega had no prior experience with proprietary trading. Instead, the team consisted primarily of employees who had significant experience designing, building and writing computer code for the suite of trading algorithms that faced ITG’s customers. These employees continued to work on the customer side of the business at the same time they were working with Project Omega, and were not physically walled off from accessing POSIT and customer data.

This matter provides a good illustration of the practical challenge many firms face in seeking to improve the culture of regulatory compliance among employees whose primary expertise involves computer coding, statistics or mathematical analyses, and who may not have a deep understanding of securities regulation. The facts here no doubt will be cited to support the recent proposals to impose a registration requirement on persons who develop trading algorithms that we blogged about here and here.