Follow Datanami:
November 8, 2016

SEC Mines Data to Spot Insider Trading

via Shutterstock

Using an internal analytics platform along with commercial data mining tools, federal securities investigators are using billions of rows of trading data collected over the last decade to spot insider-trading schemes.

The U.S. Securities and Exchange Commission (SEC) formed an Analysis and Detection Center in 2010 in response to growing market abuses. The SEC’s enforcement unit uses an in-house analytics platform along with data mining software supplied by Palantir Technologies, Palo Alto, Calif., under a five-year, $90 million contract awarded last year.

The agency also plans to launch a comprehensive database over the next year containing detailed trading information from U.S. equity and other markets.

The analytics tools allow the unit to scan databases looking for insider-trading patterns in an estimated 10 billion rows of historical trading data culled from brokerages over the last decade. The effort comes as instances of insider trading are growing almost as fast as the amount of data generated by stock transactions.

Investigators are using the historical data to discern patterns such as stock purchases ahead of mergers or U.S. approval of a new drug. Leveraging its new analytics tools, the SEC fraud unit has over the last year announced a series of enforcement actions after spotting insider-trading schemes.

Insider trading cases are often based on tips from whistleblowers or reports from “self-regulatory organizations” like the Financial Industry Regulatory Authority (FINRA) or the Chicago Board Options Exchange. A former SEC official recently asserted regulators need no longer wait for a referral when they have analytics tools and a growing database of trading information to spot insider trading.

Robert Cohen, who was named co-chief of the SEC’s Market Abuse Unit in September 2015, said earlier this year that the insider trading analysis center he oversees has so far filed several insider-trading cases. In May, for example, it announced insider charges against a professional sports gambler and a former board member of Dean Foods Co. The case also ensnared professional golfer Phil Mickelson, who later agreed to repay profits from stock purchases based on insider information.

While the SEC is ramping up its analytics tools to nail inside traders, observers not it still lacks a sufficient database of trading information. With that in mind, the agency released plans earlier this year to establish a “consolidated audit trail” that would provide a far more granular view of what happening on stock exchanges. According to reports, the new database is expected to begin gathering data on U.S. equity and options markets along with data from FINRA by the end of 2017.

Previously, Palantir has reportedly helped the SEC investigators compare trading records to information from other databases to spot patterns that might indicated insider trading. The agency is believed to be using Palantir’s Metropolis platform specifically designed, according to the company’s web site, “to easily integrate data from disparate source systems for analysis.”

Palantir also supplies data mining and database technologies to several U.S. military branches, including U.S. Special Forces Command.

Recent items:

Cognitive Platform Sharpens Focus on Unstructured Data

Machine Learning Enlisted to Fight Ransomware