You may remember that a while back the SEC decided to modernize its computer capabilities by deleting all the porn* and replacing it with algorithms that could catch other algorithms, sort of like Tron. Apparently that worked:
On Thursday, a new “analytics” division tasked with mining hedge fund data announced actions against six individuals and three hedge fund firms for alleged fraud.
“We’re using risk analytics and unconventional methods to help achieve the holy grail of securities law enforcement – earlier detection and prevention,” said Robert Khuzami, Director of the SEC’s Division of Enforcement. “This approach, especially in the absence of a tip or complaint, minimizes both the number of victims and the amount of loss while increasing the chance of recovering funds and charging the perpetrators.”
Here's how the SEC describes its, er, data mining** division:
Under the initiative — the Aberrational Performance Inquiry — the SEC Enforcement Division’s Asset Management Unit uses proprietary risk analytics to evaluate hedge fund returns. Performance that appears inconsistent with a fund’s investment strategy or other benchmarks forms a basis for further scrutiny.
Note to hedge fund managers: if you were guaranteeing "aberrational" returns, you may want to change that to a less loaded word like "abnormal" or "market-beating."
Now, far be it from me to audit the SEC. (They have a guy who does that, actually. He's pretty good.) Still, this is sort of a puzzling advertisement for the SEC's new "we now have data analytics to spot bogus outperformance, so if you make a Ponzi and tell us that it's based on split-strike conversion and/or front-running, we'll catch you" program. It's hard to tell because the SEC doesn't reveal how its algorithms caught their prey (smart!), but none of these cases scream "holy grail of securities law enforcement."
The biggest of them is Millennium Global Investments, a hedge fund that allegedly made up valuations of illiquid Nigerian and Uruguyan warrants to inflate its NAV. The fund collapsed in October 2008, losing over a billion dollars; to give you a sense of proportion, it had $844mm of assets in August 2008 so I suspect somewhere between "not much" and "a negative amount" was left over after that billion-dollar loss. Per the SEC, "As of today, the Fund's investors have not had any of their invested funds returned to them." I suspect the SEC's algos helped them figure out the warrant overvaluation (though, to be fair, I scrutinize all proposals to invest in illiquid Nigerian warrants pretty closely) but as far as early warning goes, bringing a case three years after investors lost 118% of their money seems like ... late, unhelpful warning.
Then there were two other cases, Solaris Fund and LeadDog Capital Management (solid name btw), which the SEC alleges were more or less penny-stock schemey things. It may not surprise you to learn that the lead dog at LeadDog, one Chris Messalas, had racked up a series of FINRA violations at his previous employer. Possibly the SEC is attributing to its new data-mining program anything having to do with its also-new initiative to pay attention when people rack up dozens of complaints. Not sure. But in any case the hand of the "Aberrational Performance Inquiry" is hard to detect here.
My favorite, though, is ThinkStrategy Capital Management, which was a teeny hedge fund attached to a larger fund-of-funds. The SEC is pissed about both. I suspect an algorithm helped them figure out that the hedge fund was doing this:
But on the other hand I'm not sure the algo caught the fund of funds doing this:
67. Beginning in 2005 and continuing each year through early 2008, Multi-Strategy Fund made numerous investments - totaling over $32 million - in a hedge fund called Finvest Primer ("Primer Fund"), a fraudulent enterprise managed by Gad Grieve and Finvest Asset Management, LLC ("Finvest") that was also the subject ofCommission action in February 2009.
68. ThinkStrategy conducted virtually no qualitative due diligence checks on Finvest or Grieve before making its investments. ...
69. ... Finvest and Grieve perpetrated their fraud on investors in part by disseminating false year-end financial statements with an audit opinion from a fictional accounting firm that Grieve created called "Kass Roland LLC."
70. In 2006, Finvest provided ThinkStrategy with audited 2005 Primer Fund financial statements that were purportedly certified by Kass Roland.
71. A basic effort to look into Kass Roland would have revealed that it was not a bona fide accounting firm, much less a reputable one. Kass Roland, naturally, had no name recognition or reputation. Moreover, Kass Roland had no website, its listed address in the audit report was on a non-existent street in Jersey City, and its listed phone was answered exclusively by machine. The firm was not licensed by either the Public Company Accounting Oversight Board or the New Jersey Board ofAccountancy.
Got that? Finvest was defrauding investors since 2005 by sending out financial statements audited by a made-up firm, which "a basic effort" would tell you was fake because it had no website or reputation. And these jerks at ThinkStrategy, who claimed to be protecting their investors by checking for fraud, didn't figure that out until 2008! The SEC, also charged with protecting investors by checking for fraud, found it out in ... 2009.
Well. It's nice that they're working on it anyway.
* Some of the porn. Be reasonable.
** Quick, CFA candidates: data mining is bad, right?