Not Your Granddad’s Spoof

A recent job posting by a major investment bank reads: “Basic qualifications: PhD in Computer Science, specializing in … machine learning … with extensive knowledge of big data technologies … and experience with predictive modeling, natural language processing, and simulation.”  Quantitative trading, right?  Or global risk modeling? Maybe electricity market forecasting? No, Compliance. Specifically, surveillance analytics. Twenty years ago, even ten, the Quants landing on Wall Street with their freshly minted PhDs from Stanford and MIT would have laughed. Twenty years ago, even ten, surveillance was a dreary back office affair, something that somebody in a cheap labor state did on a mainframe, if they hadn’t been laid off yet. Since Dodd-Frank, since the Flash Crash, since MiFID and MAD, since that darned book, compliance surveillance is front office with a capital ‘F’. On the trading floors of New York and Chicago, and on quieter desks from Greenwich to Boston, trading supervisors are reviewing surveillance reports and consulting real-time surveillance monitors as though their bonus checks depend on it—because they do.
Here’s why. The regulators have grown more aggressive, and grown sharper teeth—they’re now empowered to prosecute on the basis of ‘disruptive practice’ rather than ‘intent’.  Staffs are larger and regulatory actions mo
re frequent. Most importantly, money penalties have grown dramatically. Enforcement groups seem to vie with each other for bragging rights.

The regulators are better equipped now, too. They have to be, in order to analyze an ocean of market data. The CFTC’s trade surveillance system, for example, had gathered over 160 Terabytes as of June 2014, and that has likely passed 200 TB by now. Regulators are turning to sophisticated analysis to ferret out patterns of misconduct and detect market stability risks. In 2015, Scott Baugues, Deputy Chief Economist at the SEC, wrote that several SEC departments use machine learning techniques to identify likely misconduct. The old spoofs and other tricks are now easily spotted. New ones are appearing, but they’re being learned by smart analytics. Manipulate on the CME in one contract, and take on ICE in a correlated one?  Nope, a regulator’s cross-market surveillance can see it. Collude in setting a fix? Software that learns social media relationships may detect it. So what should the compliance chief who wants to protect her firm from a business-busting fine do?  The first thing is to wake up to how dramatically the surveillance landscape has changed—it’s not your granddad’s spoof anymore.

chart


References:
Job posting: http://www.goldmansachs.com/a/data/jobs/27931.html
CFTC Enforcements and penalties are from the CFTC Enforcement Division annual results archived on the CFTC Press Room web site.
CFTC Surveillance collects 160 TB: https://fcw.com/articles/2014/06/03/cftc-mulls-retooling-market-surveillance.aspx
Scott Baugues comment: http://cfe.columbia.edu/files/seasieor/center-financial-engineering/presentations/MachineLearningSECRiskAssessment030615public.pdf

thanks for reading.
Dermot Harriss

About Louis Lovas

Director of Solutions, OneMarketData, with over 20 years of experience in developing cutting edge solutions and a leading voice in technology and trends in the Capital Markets industry.
This entry was posted in Algorithmic Trading, Analytics, Big Data, Cloud Computing, HFT Regulation, Tick database, Trade Surveillance. Bookmark the permalink.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s