Very Fast Insider Trading

by on February 2, 2013 at 9:49 am in Economics | Permalink

On January 31, 2013, approximately 400 milliseconds before the official release of the EIA Natural Gas Report, trading activity exploded in Natural Gas Futures and ETFs such as UGZ, UNG and BOIL. Now that the Feds have stated (as claimed by a recent WSJ article) that they don’t think there is merit in prosecuting people who get news information earlier than others by milliseconds, is it any wonder?

More here. Previous MR posts on high frequency trading.

nathan w February 2, 2013 at 10:14 am

I thought insider trading, etc., was highly illegal. What difference does it make it 400ms or a month, if the outcome is a large volume of trades clearly based on information that people shouldn’t have had.

Bill February 2, 2013 at 10:30 am

Please provide a link to the WSJ article. I do not believe that the SEC would state that if you acted on inside information you would not be prosecuted, even if you “acted” a millisecond before the release.

There is no link to the claim in your post.

Yours Truly,

Doubting Thomas.

Nylund February 2, 2013 at 10:59 am

I’m hesitant to do work for people who are too lazy to do it themselves. Considering it took about 5 seconds using Google, I’ll let it slide:

http://online.wsj.com/article/SB10001424127887323539804578262280735396300.html

The relevant passages:

“Investigators recently decided not to file any criminal charges, said people familiar with the matter….A key issue, one of the people said, was whether the government could prove in court that a time advantage for a trader of a sliver of a second—as little as a few thousandths—was enough to conduct profitable trades on confidential information.”

BC February 2, 2013 at 12:55 pm

And, the sentence before that passage is, “…the investigators decided against filing charges because they couldn’t link the [trading] pattern[s] to specific actions by media companies, people familiar with the probe said.”

After reading the article in its entirety, I did not get the impression that the Feds “don’t think there is merit in prosecuting people who get news information earlier than others by milliseconds.” Rather, there is a complicated protocol for releasing data with evolving technology from multiple sources, and the Feds are still trying to figure out where the vulnerabilities are and to what extent traders may be exploiting those vulnerabilities: ” ‘There is a vulnerability there, but agents just can’t prove that it was being used illegally,’ said a person familiar with the investigation.”

Bill February 2, 2013 at 9:18 pm

Nylund,

BC was correct. After reading the complete article, which, by the way, failed as a matter of proof (that the disclosure was inadvertant) rather than policy, I think that the information should be released by the agency, and not embargoed for release by an entity which failed to have sufficient internal controls to release it on time, which means in this case, in time with the atomic clock, as was required. Second, what was involved here was a criminal standard. The rules are not changed AT ALL contrary to the what the post suggests.

it should be released by the agency. Simple solution.

Alexeisadeski February 2, 2013 at 10:58 am

More likely that the 400ms traders were responding algorithmically to other, earlier inside and speculative trades, no?

Every major announcement of every type has a flurry of trading activity prior. But may be rumors, speculation, mob mentality. Separating inside information from this is quite challenging.

HFT Trader February 2, 2013 at 3:08 pm

Exactly. Activity will always pick up immediately preceding an announcement, even before no information is released. Frequently this is because many people are positioning themselves to either take advantage of a view on the event or scale down risk.

I very much doubt that if someone truly had inside information they’d wait till the second before the announcement. You simply can’t scale up that much liquidity in such a short period of time. You’d make a lot more money by building up a large position, which takes time.

Rahul February 2, 2013 at 11:05 am

What surprises me is that the EIA Natural Gas Report has any meaty info relevant to trade strategies. Those reports are usually summaries and not novel insights.

HFT Trader February 2, 2013 at 3:06 pm

I used to sit on a (non-HFT) energy desk for a large hedge fund. One of the strategies we used to run was just a systematic strategy where we’d buy or sell depending on the weekly EIA numbers. That is we’d start trading after they were released and keep trading in that direction until the end of the session. Then liquidate the position on the next session. It was one of the most profitable strategies on the desk.

In short the market would trend for the rest of the day in the direction of the announcement, even hours after it was announced. It’s not uncommon to see phenomenon like this in the context of large announcements having an effect over a period of hours to several days. Look up relevant research on single stock biases post-earnings announcement as well as market-wide biases post-Fed announcement.

Corporate Serf February 3, 2013 at 8:59 am

These storage numbers are extremely important, for momentum type trading, which of course has implications if you are just hedging a derivative position based on it.

Re- positioning before the numbers, on the energy options desk I used to sit at, the traders would put in limit orders, but not actual trades. From my memory, and an admitted a qualitative feel, it seemed to me that the number of incidents of early trading slowly increased. I am sure, a lot of it was fat fingers, given the increasing attention the NG and CL futures started receiving from non-energy type traders; (also, given that it wasn’t always the case that the direction of these early trades was correct; though that might easily be some other algorithm trying to hit the limit orders in the book) but one can never preclude early release, whether malicious or inadvertent.

I am glad these issues are getting some attention, but am pessimistic about the investigative effectiveness of the regulators.

Stephen Long February 2, 2013 at 11:39 am

How is it clear that this is just one person reacting to the news and not just the entire market? It is no surprise that there is an explosion of trading immediately after the weekly eia data is released. It is also important to note that the number came in at a draw of 194, which was 11 bcf under expectation, a fairly large surprise. Given this fact trading volumes would be expected to be greater than normal for Thursday, although they were still very high. I find that Nanex and Zerohedge behave as if they are conspiracy theorists, if they throw enough crap at the wall something eventually will stick.

There has been some discussion regarding high frequency trading firms attempting to trigger stop orders during times of market illiquidity (“Banging the beehive”), seconds before the EIA number is released definitely falls into that category. It is very possible that whatever firms have been doing this attempted to and that the number came out while the algorithm was running, the attempt to trigger stops on the upper limit would be in that quick pop you see after the initial sell off. It is also possible that this attempt was so close to the number that other high frequency firms piled on the move immediately.

Rahul, the EIA number is highly relevant to the natural gas market… You should expect a significant move every Thursday @ 10:30 est as a result of it.

Last, the more interesting move was the rip back up later in the day. I am interested in what drove that….

Mark Thorson February 2, 2013 at 11:54 am

Is there an assumption that nobody can digest an IEA report and make trading decisions based on it within 400 ms? It’s possible you’re only looking for one number in that report, and that the number occurs at a predicatable place within the report. If you had trades prepared merely waiting for a click on a button and a fast way to pop the number out of the report, I think you could do it in 400 ms even with a human in the loop.

Doubt February 2, 2013 at 11:59 am

One would have to be a pretty fast reader to identify and act on the number 400ms BEFORE the report was released.

zbicyclist February 2, 2013 at 12:01 pm

It’s 400 milliseconds BEFORE. So, if you had inside information, you might have had it for some time. You may have decided that your best strategy to make money and evade prosecution was to make trades almost simultaneously (400 milliseconds ahead), figuring (evidently correctly) that even if somebody noticed they would decide it would be too difficult to try and prosecute.

Ray Lopez February 2, 2013 at 12:28 pm

BEFORE, Before, before. yes that is key. A theory of mine (that I cannot prove) is during the week in September 2008 that ended with the Friday where Hank Paulson asked for bailout money from the US Congress, with the plea that the world would end otherwise, the markets froze up because they knew (via inside information) that Paulson was going to ask for a bailout. Why trade toxic assets at a discount if DC will bail you out at 100 cents to the dollar? Inside information is rife–it’s the entire foundation of ‘mo-mo’ investing and most hedge funds.

dead serious February 2, 2013 at 12:44 pm

What this means, to nobody’s surprise, is that the financial industry government watchdog groups are incompetent and toothless.

Douglas Knight February 2, 2013 at 1:05 pm

How hard is it to prove that someone made money on this? Just add up the trades. It looks to me like there were 50 trades on 10 exchanges. How hard would it be to subpoena the records and find out who made these trades? If they were all the same party, that’s pretty damning. If they were all different parties, especially if one party traded in both directions, that suggests the “bang the beehive” theory.

Why doesn’t Nanex at least say how much money the first trade made and how much could have been made if all the trades were by the same party?

ewg February 2, 2013 at 1:16 pm

Although the article discussed trades that occurred in natural gas ETFs on various stock exchanges. I want to throw in $0.02 and say that there is no insider trading in commodities markets; i.e. it is perfectly legal to use insider information to trade. (Presumably it is still illegal to acquire the insider information.)

If I were privy to insider information about an upcoming government report, I wouldn’t bother to trade the ETFs at all. Why expose yourself to possible SEC legal action? My point being is that there is no reason to trade the ETF when you could legally trade the future’s contract.

At the link, I can’t tell if there were trades 400ms before the report on the future’s contract, only on the ETFs. I would like to see what happened the the natural gas future’s contract, basically the equivalent of the first plot.

Ray Lopez February 2, 2013 at 1:31 pm

“I want to throw in $0.02 and say that there is no insider trading in commodities markets; i.e. it is perfectly legal to use insider information to trade. (Presumably it is still illegal to acquire the insider information.)” – does not seem to completely compute. They made a movie in 2009 with Matt Damon, “The Informant” about price fixing, and I would assume trading is part and parcel with this crime.

Cliff February 2, 2013 at 2:37 pm

Huh? Price fixing and insider trading could not be more different.

Alex Godofsky February 2, 2013 at 1:17 pm

This may be a stupid question, but how confident should we be in their claim that the trading was 400ms before? Seems like that relies on the clocks of two (presumably) separate computer systems being precisely synchronized. I know the computer network at my office is actually off by two entire minutes. Is drift of a few seconds that implausible?

Stephen Long February 2, 2013 at 1:20 pm

I agree 100%, that was my first immediate thought.

Stephen Long February 2, 2013 at 1:22 pm

The EIA has botched the release before as well, its been delayed and released early. The fact that it came out 400 ms before its scheduled time is definitely a possibility.

GSW1200 February 2, 2013 at 2:30 pm

+1

Douglas Knight February 2, 2013 at 2:32 pm

The Nanex graph shows two spikes in trading activity, 400ms apart, which is not what I’d expect if there were one early release. But the later spike, labeled official, is smaller, which is not what I’d expect if one party had early access. But maybe my expectations are wrong. Maybe the one party moved the market, so there was little additional movement at official release.

alexh February 2, 2013 at 5:39 pm

If you office is off by two minutes, best guess is that no one there cares at all. At all. Even basic windows desktops have an option for “internet time synchronization” (NTP), which if pointed to an accessible server (try time.nist.gov) should give you accuracy well better than a second. Basically by checking a box!

At worst case, financial enterprises will install at least one GPS server somewhere, use NTP, and get single-digit milliseconds (or better) of globally-comparable accuracy. That’s going to
be the least professional setup one should see. If one cares (and _most_ exchanges do) you can without much extra work or cost get much tighter than one millisecond
(or with some real work, much much tighter).

Issues do happen, but with everyone involved in this story I’d expect a 400ms clock skew problem to be (a) very very rare, (b) noticed more or less immediately, and (c)
acted on and fixed as quickly as possible. Basically, that doesn’t count as a small discrepency or even close.

JWatts February 2, 2013 at 1:19 pm

Perhaps someone got their quantum computer on-line? ;)

Kid Dynamite February 2, 2013 at 1:42 pm

it seems like lots of people didn’t bother to read the article (which itself doesn’t answer all of the questions, of course)..

the trading occurred 400ms before the OFFICIAL release, because Bloomberg and others had figured out a flaw in the release process LAST SUMMER (which they told the Dept of Commerce about). So, essentially, the data was released 400ms early:

“That happened, the people said, after Bloomberg tested its computer system at an embargo procedure known as a “lockup” and found a flaw that could allow reporters to deliver data before the end of the embargo period. “We discovered a flaw in the Department of Commerce’s lockup system during a testing period sponsored by the department,” said a Bloomberg spokesman. “We alerted Commerce to this flaw when we saw it had potential to be exploited by others to beat their embargo during an actual lockup. We suggested solutions to secure their system and they thanked us for alerting them to the issue.””

then there’s this, which drives me crazy:

“SEC and FBI investigators have harbored suspicions since 2007 that some traders may have gained access to information slightly before it was public, according to people familiar with the probe and an audit of Labor Department systems.”

no – they got the information *when* it was public (thanks to the data providers) – which was slightly before the rest of the public got it (because these traders want the super fast feed and pay for it) through the general release. The traders aren’t stealing the information – although there is some sort of insinuation here that the data providers (bberg et all) ARE stealing the information. Either way, this isn’t a story about high frequency traders trading on inside information: somebody will ALWAYS have the data first: http://kiddynamitesworld.com/someone-will-always-have-the-data-first/

this MIGHT be a story about the data providers stealing information – that remains to be seen.

oh by the way, this is also ridiculous:

“A key issue, one of the people said, was whether the government could prove in court that a time advantage for a trader of a sliver of a second—as little as a few thousandths—was enough to conduct profitable trades on confidential information. ”

of course it’s enough time. THAT is where the high frequency traders come into the discussion.

Dan Weber February 2, 2013 at 4:43 pm

Here’s my question:

If someone is eager to trade 400 milliseconds before the information is released, why would you volunteer to be on the other side of the table? Either you think he’s a sucker and you are taking advantage of him (for which I have little pity), or he has advance information and is going to be taking advantage of you, in which case you stay the hell away.

Shane M February 2, 2013 at 9:26 pm

+1. My thoughts exactly. The biggest news of the week comes out in 400ms but I’ll take a big trade right now instead of say, 5 minutes ago.

Corporate Serf February 3, 2013 at 9:07 am

Indeed. As far as I can tell, your argument also applies if the early trader got the information a day early, rather than 400 ms early. Correct?

Comments on this entry are closed.

Previous post:

Next post: