An indictment unsealed in federal court in Boston serves as a stark warning to lawyers and their clients of the increasing regulatory risk posed by the Securities and Exchange Commission’s use of advanced analytics to detect illicit trading activity.

On Sept. 12, the U.S. Attorney’s Office announced that Michael Bressman had been arrested and charged with securities and investment advisor fraud. The government alleged that the New Jersey broker had engaged in a “cherry-picking” scheme to obtain $700,000 in illicit trading profits over a six-year period.

The SEC filed fraud charges against Bressman in a parallel action in New York federal court. In announcing the enforcement action, the SEC highlighted that the alleged fraud had been uncovered through the use of data analysis that revealed “suspicious trading patterns” on the part of the defendant.

“We will continue to develop and use data analytics to root out cherry-picking and other frauds,” said Joseph G. Sansone, chief of the SEC Enforcement Division’s Market Abuse Unit in a statement accompanying the agency’s announcement.

One attorney who views the SEC’s emerging data analytics capability as a serious regulatory risk for clients is Neil T. Smith.

“They’ve been using analytics for a while, but they’re fine-tuning it now,” said the Boston lawyer, who practices in the fields of government investigations, securities enforcement and white-collar defense.

According to Smith, the SEC is moving toward an aggressive data analytics-driven enforcement regime whereby the agency is able to compile and present comprehensive trading analyses demonstrating a potential violation, prior to even opening a formal investigation against an entity.

“They’re not shooting first and asking questions later, but they are coming in ready to charge,” Smith said. “They’re saying, ‘We don’t believe you can possibly refute the data, so you either settle with us or we’re going to charge you.’”

Smith predicted that, as the SEC obtains more favorable outcomes through enhanced analytics, it will become even more inclined to use such tactics.

False positives

Ian D. Roffman, a securities enforcement lawyer in Boston and former senior trial attorney for the SEC, pointed out that the agency uses analytics both to monitor trading activity and examine public company financial statements.

And though securities attorneys view large investment firms as having the resources to deploy their own internal analytics programs to police themselves, lawyers like Roffman are concerned that smaller companies may not have the means to avoid “false positives” when the agency’s data analytics trigger a costly SEC investigation.

“Everyone agrees that the [SEC’s] use of data analytics is a good thing to the extent that it is able to identify bad actors and remove them from the industry,” Roffman said. “But data doesn’t always tell the whole story. Any analytics tool may have built-in assumptions that may or may not be valid.”

According to Roffman, the SEC needs to be careful it doesn’t take advantage of smaller companies by “jumping to conclusions” too quickly. He added that the agency should be sensitive to the costs that data analytics tools impose on “good-faith actors” who want to ensure they’re not caught up in the search for bad actors.

“It needs to be open and transparent about its data analytics so that people facing an investigation have a fair opportunity to understand the data the government is looking at and respond to it,” Roffman said.

In order not to be swept up in the SEC’s net, Roffman said that honest market participants need to have a data analytics component as part of their own internal compliance program. Such a program should allow the company to look at trading patterns in real time in order detect any problems that need to be addressed.

Home-grown program

Analytics came to the fore of enforcement activity early in 2011 with the creation of a detection center in the SEC’s Market Abuse Unit. The center is staffed with specialists with trading experience who use technology to analyze “billions of lines” of trading data to investigate and build insider and abusive trading cases.

At a technology summit this past spring, former SEC Chairman Michael S. Piwowar described one of the commission’s key technological enforcement tools for detecting insider trading and market manipulation activities.

The Advanced Relational Trading Enforcement Metric Investigation System was developed in-house by commission staff. Piwowar explained that ARTEMIS combines historical trading and account holder data with other data sources “to enable longitudinal, multi-issuer, and multi-trader data analyses.”

According to a Reuters report, though the SEC does not have a “direct feed” of market trading data, it has the capacity to mine billions of lines of “Blue Sheet” data of trades executed by brokerages. Blue Sheet data files contain both trading and account holder information and provide the SEC and other regulatory agencies the ability to analyze a firm’s trading activity.

Investment firms are required to provide “complete, accurate and timely” Blue Sheet data in response to regulatory requests, allowing regulators to better identify insider trading schemes and other fraudulent activity. The ARTEMIS program analyzes historical patterns and relationships from the Blue Sheet data.

For example, in the insider trading context, ARTEMIS flags traders who exhibit an unusual pattern of making judicious trades in advance of the release of company news that affects stock prices.

“We surveil the trading in the securities markets to identify patterns of suspiciously profitable trading,” Sansone told New England In-House.

For example, Sansone said, the SEC in May filed insider trading charges in federal court in New York against investment banker Woojae Jung. The SEC alleged that Jung used sensitive client information in order to trade in the securities of 12 different companies prior to the announcement of “market-moving” events.

The SEC further alleged that the defendant used an account held in the name of a friend living in South Korea to place the illegal trades, which generated profits of approximately $140,000.

“Like others before him, Jung’s alleged scheme failed when our data analysis uncovered the account’s suspicious trading pattern and, despite Jung’s attempts at evasion, traced the trading back to him,” Sansone said in an agency statement announcing the charges.

“Everyone agrees that the [SEC’s] use of data analytics is a good thing. But data doesn’t always tell the whole story. Any analytics tool may have built-in assumptions that may or may not be valid.”

— Ian D. Roffman, securities enforcement lawyer

Going after the little guy?

For years, the mining of data has been common in the SEC’s prosecution of the big insider trading cases. But New York attorney Meghan K. Spillane said she’s detected a recent trend in which smaller traders are also being caught in the agency’s net.

“It’s not limited to these big players and these big events,” said Spillane, who practices securities law and white-collar defense. “The SEC also pursues cases against relatively small, individual traders that make relatively minor profits.”

According to Spillane, the SEC’s computers and programs have “gotten smart” about the way individual market participants typically trade, allowing the agency to better detect anomalies and an individual’s break from his or her normal trading patterns.

“Everyone should be aware that the SEC currently has the tools to look at an individual trader’s conduct,” she said. “The SEC is not simply being reactive to a big event or a tip; it’s acting proactively to detect activity that years ago may have been missed.”

According to the government, analytics helped reveal that defendant-broker Bressman allegedly made certain trades, placed them in a holding or “allocation” account, waited for a short period to see how the stocks fared, then “cherry-picked” the profitable trades, which he transferred from the allocation account to his own account or the account of two family members.

Trades that turned out to be unprofitable allegedly tended to land in the accounts of his customers.

Hot spot for enforcement?

With so many financial service companies in Massachusetts, it should come as no surprise that the U.S. District Court here has become a hot spot for SEC enforcement featuring the use of analytics.

In January 2017, the SEC filed fraud charges against Massachusetts-based investment advisor Michael J. Breton and his firm Strategic Capital Management. The SEC alleged Breton defrauded clients out of approximately $1.3 million in another cherry-picking scheme detected by data analysis.

With the announcement of the charges, the agency further announced that Breton had agreed to be banned from the securities industry.

After Breton later pleaded guilty, U.S. District Court Judge Allison D. Burroughs sentenced him to two years in prison and two years of supervised release. The judge further ordered the defendant to forfeit $1,326,696 and to pay restitution in the same amount.

Sansone told New England In-House that cherry-picking cases are a good example of how valuable data analysis can be used as an enforcement tool.

“We can use the analytics to show that a hugely disproportionate number of the profitable trades ended up in the investment advisor’s own account, as opposed to his clients’ accounts, and to show through statistical analysis that that was not an accident or coincidence,” Sansone said.

Data analysis also plays an important role in generating evidence necessary to build a case against a defendant, even when it may not initiate the case, according to Sansone.

He pointed to a case in which U.S. District Court Judge Rya W. Zobel entered final judgment in February, after the defendant agreed to settle with the SEC. In SEC v. Amell, the SEC alleged that a Massachusetts-based portfolio manager at an unidentified “major asset management firm” diverted $1.95 million to his personal brokerage account from a fund over which he had trading authority.

According to the government, Kevin J. Amell conducted a “matched-trades” scheme in which he pre-arranged the purchase or sale of call options between his own account and the brokerage accounts of the fund at prices that were disadvantageous to the fund and advantageous to him.

The final judgment required the defendant to disgorge $1.95 million realized from the fraud. The defendant had pleaded guilty in a parallel criminal case in which he was sentenced to 18 months in prison.

Sansone said evidence generated by the SEC’s data analysis experts was critical in “marrying” the trading that was done on behalf of the investment firm with the trading that was done in the defendant’s individual account to show how he “systematically” benefitted from trading against his clients.

“The power of this evidence helped the government secure a prompt guilty plea and favorable settlement in the SEC case,” Sansone said.