Comment by elliekelly | Hacker News Reader

Comment by elliekelly | Hacker News Reader

Parent Full thread

It’s part of their new “EPS Initiative” where they use an analytics tool called ARTEMIS (I forget what the acronym stands for) to assign a risk score to certain transactions and traders. It’s interesting because although the SEC has been pretty tight-lipped about the data inputs it appears there is some degree of “who-you-know” factored into the risk score. So two people can make the exact same trade at the exact same time but if I have a close friend or family member affiliated with Netflix and you don’t then my trade will receive a higher risk score.

It’s part of a broader change in the way the SEC approaches insider trading. Previously they took an “issuer based” approach where a big pop or loss in $ABC triggered a look at all the trades in the right direction prior to the announcement. Now they take a “trader based” approach where they look at person’s pattern of activity over time.

So if your spouse or sibling works in M&A and tips you off to potential deals so you can make (relatively) small investments on the inside information the scheme will be flagged and investigated much more quickly.

How does a system created by the SEC detect that person A is a (close) friend of person B, with accuracy? Family I can understand, if you share the same last name etc, but friends?

This is the most important question IMO.

However, I'm not sure they do have that information. I would speculate that the number of people who make large option bets before earnings correctly on the same stock over multiple quarters should be zero and if it is very small, it is worth investigating. If they see that a subset of traders is 3 people with Korean names, that would immediately suggest that something is wrong...

What’s really interesting is I believe at least one commissioner (or perhaps staff member?) has made comments hinting that it works the other way, too. It can identify “related” transactions executed by other seemingly random people and, when they look into it, they find out the person who executed the trade does indeed know the original person under investigation.

Dont really know enough about this to do anything other than wild speculation but Id lean towards something like: If a trade crosses a certain risk score threshold it would trigger some kind of preliminary investigation looking at limited information. If the preliminary investigation determines the trader has relationships with a company insider, its then added to the metrics and the risk score may ramp up to a level which would trigger a wider investigation.

But honestly, it’s much more likely that it was simply the timing of the trades and the earnings reports multiple times over a couple years raised some eyebrows.

They have lists of all your phone contacts (if you uploaded them to Apple, Google or socials), your socials friends/followers, the people you call and text (via cell service providers)... These all get shared interagency. It's not like the SEC is collecting it themselves.

They could have their own little PRISM.

ARTEMIS stands for the Advanced Relational Trading Enforcement Metrics Investigation System.

Also Artemis was the goddess of hunting in Ancient Greece.

And perhaps more interesting to WSB folks, Artemis is also the goddess of the Moon.

I don't think SEC has any encrypted messaging reading capabilities but average guy suddenly making regular killing on the market with the same stock does turn on red flags.

And then it's a matter of connecting the dots - most often the dots between the "lucky" guy and someone within the company.

Done.

I'd hazard a guess that you don't even need to make a killing or do it suddenly--every insider scheme I've heard about involves getting company performance data a few days before it is publicly released at an earnings call or a 10k (or notice of an M&A a few days before it is announced).

If you consistently make directionally-correct trades (no matter how big) on one particular symbol (or derivatives of it) within +/- n days of earnings, that probably raises your risk score. Probably even more if your trades are always the same day of week before earnings or same number of days before earnings.

I was working overseas once, and a cousin I’ve never met coincidentally purchased some stock in my employer. It wasn’t a huge order, and there was nothing especially strange about the transaction (no massive gains or anything). But their local version of the SEC investigated it and got in touch with my company about it.

I’d be surprised if the SEC wasn’t proactively monitoring this sort of thing on some level.

That's all well and good, but statistical anomalies and potential personal ties to insiders don't prove cases, especially criminal cases. Unless they have the actual texts from whatever "encrypted messaging apps" they were using, they won't people to prove these allegations. If you could go to jail merely for being suspiciously lucky in the markets while knowing corporate insiders, lots of hedge fund managers would be in prison right now.

Correct. That is why they use this circumstantial evidence to get a search warrant to confiscate your phone and other devices, read your email and other messages, and then question you until you, or one of your co-conspirators, breaks.

...which includes lying to you and telling you that the other parties have already confessed, and that you'll be the one holding the bag in jail if you don't confess as well.

A few hours of this, and unless you have the foresight and tenacity to demand a lawyer, most people break. Only one of them needed to break in this case.

How do they know whose friends are friends with whoever else?

I'm certain it's easier than basically any point in human history to obtain such information from: intelligence agencies, social media companies, data brokers/advertisers, etc.

Your public friends list on LinkedIn, Facebook, etc?

Reply on news.ycombinator.com