Follow Datanami:
August 19, 2014

U.S. Enlists Big Data to Fight Securities Fraud

The U.S. Securities and Exchange Commission (SEC) moved to reform its enforcement practices after missing the Bernard Madoff Ponzi scheme and other financial fraud by creating an Office of Market Intelligence in 2010. The office was intended to make use of new intelligence tools like big data analysis to improve securities enforcement.

So far, the results of the SEC efforts have been mixed, according to the author of a recent law review article focusing on how the office has leveraged big data. The author, Geoffrey Rapp of the University of Toledo College of Law, concludes that the SEC intelligence unit has succeeded in some limited areas, including the recent suspension of trading in “microcap shell companies.”

But the SEC market intelligence office has also come in for criticism from the agency’s own inspector general.

SEC enlisted the MITRE Corp. to help develop the financial data analysis platforms used by the Office of Market Intelligence. An artificial intelligence upgrade was planned to help fraud investigators connect tips and complaints about financial fraud that may have not appeared to be connected. The technology uses statistic analysis to trace the connections between tips.

Rapp wrote that MITRE’s approach suggests that the SEC office is “embracing the Bayesian approach pursued by federal intelligence activities in the defense and security contexts.” He added that this approach “in the financial fraud context would look at the timing of various activities and focus both on unique indicators (those that are only associated with financial fraud) and on visible indicators (those that can be measured).”

He suspects the SEC office is using a variety of data mining techniques such as logistic models, neural networks and decision trees to more quickly detect fraud in U.S. financial markets.

A daunting challenge for the market intelligence unit is the sheer volume of data generated in the financial markets under the SEC’s authority. “Without a mission tailored around enforcement targeting priorities,” [the Office of Market Intelligence’s] effectiveness will be undercut,” Rapp warned.

To succeed, he said SEC’s big data effort must “embrace the notion of an intelligence cycle, and that intelligence cycle should specify the various entities responsible for each step in the intelligence process.”

Rapp added that the SEC office must develop metrics that go beyond how quickly it processes data to develop measurements of progress based on whether it enforcement activities are keeping financial markets operating within existing securities laws like the 2011 Dodd-Frank financial regulatory reform law.

Among other things, the Dodd-Frank law creates bounties for tips on security law violations. The market intelligence unit is responsible for following up on leads generated by these tips. It was the SEC’s failure to respond to repeated tips and a significant amount of credible evidence that allowed the Madoff scandal to fester for years before his Ponzi scheme collapsed under its own weight.

For now, Rapp concluded, the jury is still out on whether big data can help prevent the next Madoff scandal.

Recent items:

Big Data: A View From Wall Street

IBM Flushes Out Fraud with Big Data Analytics