Amazon.com, Google, and Chinese Internet searcher Baidu are among the companies with massive computing capacity that you might expect to be deploying small armies of computer scientists intent on developing potential uses for artificial intelligence. But now some perhaps less obvious players are getting into the machine learning game: financial regulators.
In an October speech, for example, Scott Bauguess, deputy director (and current acting director) of the Securities and Exchange Commission’s economic and risk analysis division, said the regulator was leveraging topic modeling and sentiment analysis to “map these topic and tonality signals into known measures of risk — such as examination results or past enforcement actions — using machine learning algorithms.”
Bauguess, who has a bachelor’s and a master’s degree in electrical engineering from the University of Illinois and Arizona State University, respectively, can speak with authority about such techniques, which “are allowing analyses of previously impenetrable information sets — for example, those without structure, such as free-form text.” Dealing as it does with free-narrative documentation from the firms it regulates, the SEC is taking a “machine learning approach to behavioral predictions, particularly in the area of market risk assessment, which includes the identification of potential fraud and misconduct,” Bauguess said.
The SEC and Bauguess stand out as, dare I say, regulatory technology innovators. And they are not alone. While regtech has recently attained buzzword status, it has manifested mainly in the context of systems and services that help with firms’ regulatory compliance and reporting. For example, NEX Group, the London-based automated trading and data services company formerly known as ICAP, seized on the trend last fall when it acquired Abide Financial, a specialist in regulatory reporting technology. And in January capital markets software provider Risk Focus moved its Report-it compliance product into a new unit, RegTek Solutions.
Regulators, meanwhile, are blazing a regtech trail of their own, turning on its head the long-held assumption that supervisory agencies cannot keep pace with the deep-pocketed firms they oversee. Several — notably, the U.K.’s Financial Conduct Authority, the Monetary Authority of Singapore, and the U.S. Office of the Comptroller of the Currency — have signaled support, and in some cases outright encouragement, of entrepreneurship in fintech.
Although a February report by the International Organization of Securities Commissions suggests that fintech adds a layer of complexity to such conventional supervisory responsibilities as investor protection and financial stability, the multinational regulatory group affirms that fintech and regtech can improve compliance oversight with data analysis tools and software. “Regulators also may explore leveraging new compliance software and surveillance tools,” the group said in the report.
The SEC began to do exactly that in reaction to the 2008–’09 financial crisis and the 2010 flash crash — systemic breakdowns that regulators didn’t see coming. Hedge fund industry veteran Gregg Berman joined the SEC’s division of trading and markets and spearheaded development of the Market Information Data Analytics System (MIDAS), which seeks to improve transparency to analyze trends and prevent crashes. The more ambitious Consolidated Audit Trail (CAT), an idea first floated in 2010 to give the SEC a complete database of market activity, is finally getting off the ground following the awarding of a contract in January to Thesys Technologies, the same trading technology company that worked with Berman on MIDAS. (Berman left the SEC in 2015 and is currently director of research — market integrity, monitoring, and surveillance at Citadel Securities.)
CAT will be “the largest data repository of securities trading activities that has ever existed,” predicts SEC commissioner Kara Stein, who has been advocating, thus far to no avail, the creation of an Office of Data Strategy to coordinate data and analytic initiatives dispersed throughout the agency.
The SEC took another leap toward parity with the private sector in 2015 with Kx Systems’ kdb+, a high-performance database more typically used in trading and risk management by the biggest Wall Street banks.
Elsewhere in Washington, the Financial Industry Regulatory Authority is developing machine learning programs for transaction surveillance. And the Commodity Futures Trading Commission has turned to outside vendors Neurensic and Vertex Analytics to do the same, Bloomberg reported this month. Acting CFTC chairman J. Christopher Giancarlo, formerly an executive of tech-focused interdealer brokerage GFI Group, leaves little doubt about his vision. “We need to make the CFTC a 21st-century digital regulator and no longer an analog regulator,” Giancarlo told Risk Desk newsletter in January. “We need to partner with fintech companies to raise our game.”
These watchdogs may be budget-constrained, but they are not to be sold short.
