It is not very often that a financial story comes along that requires referencing the Einstein’s theory of relativity to explain, but we have finally come to this.
One week ago, the Fed made an important surprise announcement — that it would not be reducing its bond buying program. The result of this announcement was that the price of stocks, bonds, and gold skyrocketed. In order for markets to be fair, everyone must receive this information simultaneously. However, in the crazy world of relativity, there is no such thing as simultaneous. And somebody used this fact in order to make a lot of money.
This is how it went down. Anyone who has watched the “Back to the Future” movies (or the movie “Trading Places”) knows that if you know in advance which way the markets are going to move, you can make trades that will make you rich beyond your wildest dreams. But in this case, you don’t need to own a DeLorean that can travel in time. After all, it is already the future somewhere else.
When the Fed makes a market moving announcement like this, they release the news at a very specific time. In this case 2pm in Washington DC, “as measured by the national atomic clock“. Now here is where it gets tricky. According to the theory of relativity, information cannot travel at faster than the speed of light. So it should take at least 3.2 milliseconds for this information to get from DC to Chicago (a millisecond is 1/1000 of a second, and light travels around 300 kilometers in one millisecond). But someone cheated the speed of light, and received this information in Chicago at exactly 2pm, which to them is 3.2 milliseconds early. With modern high speed stock trading systems (available only to the financial elite of course), that was just as good as having their own time machine.
Receiving the information a few milliseconds before anyone else in Chicago allowed them to make several hundred million dollars in trades before anyone else. One to two milliseconds after 2pm (Chicago relative time), someone made (according to estimates) around $600 million in trades.
Why should you care about this? Because the money made by high-speed trading systems taking advantage of quirks like this is money that is siphoned away from everyone else, including you. It is money that your retirement nest-egg investment doesn’t make. And since billions of dollars are being spent on high-speed trading systems (including special high-speed light-carrying fibre optic communication lines, advanced computers, programming talent, and other resources that allow the ultra rich to execute an equity trade in around 10 microseconds), you know that they are taking a lot more than billions of dollars from people like you. In 2010, high-speed trades accounted for around 75% of all US equity trades. Which is a big reason why almost all the money made in the last decade has gone to the top 1%.
17 Comments
Who is ‘them’ and why couldn’t they be prosecuted for insider trading as clearly they had the information before the announcement?
Very interesting, so someone already knew about the announcement (well in advance). Just to avoid suspicion, they were told to do it at 3pm (2pm chicago). Is that what happened?
Are you sure about your theory? Could someone just have been making a risky bet, which is what day traders do. How long after 2pm did Bernanke make his statement about bond buying. The Fed Chairman usually speaks for a few minutes at least before getting to the point. Was it published in a document and posted online? It would be impossible for a human to speak any word in 3.2 ms.
I think people are jealous that someone bet alot of money and won the bet. Now if they actually did know ahead of the release then someone inside the Fed (which has no government oversight) sent the info early and in that case they should be prosecuted. They probably wouldn’t though because they are likely also on the big doner list.
My question is, why does day trading even exist? Why does anyone need to trade more than once per day?
There’s no trading at all on weekends, and the economy doesn’t collapse. A few decades ago, the stock market was only open 4 days a week. That didn’t cause a problem either.
75% is the total number of trades, and doesn’t take into account the money made. High frequency trading algorithms monitor the market and attempt to make money from small anomalies, so they don’t actually make that much money overall (the last figure I heard for one firm was 10% of their equity revenue). This is likely what happened here, a small anomaly was spotted and the algorithm took advantage of that. Whether this is a safe thing to be doing is a different matter entirely.
Here’s an update:
http://www.ft.com/cms/s/0/eb5953b8-2545-11e3-b349-00144feab7de.html#axzz2fw0enoBt
looks like some news organization may have leaked the information ahead of the scheduled release.
In this release it talks about Reuters regularly releasing data ahead of schedule until asked to stop by the NY Atty Gen
http://www.usatoday.com/story/money/business/2013/09/24/federal-reserve-taper-leak/2861669/
all interesting none the less. Unfortunately, the gov’t is not likely to pursue this with any gusto so as to not p@#$ off any potential donors.
News organizations are given the information early, if they promise not to release that information before 2pm eastern time. Since the news was released in Chicago *after* that time, that complies with the letter of the agreement, but not the spirit. It was still before anyone else in Chicago could have received the news, and long enough to make some profitable trades.
None of this has anything to do with how long it takes to speak or the length of Bernanke’s introductory remarks. The “news” is released electronically. Companies pay lots of money to get the information instantly, so they can profit from it.
tl;dr You guys asked lots of questions that would have been answered if you read the linked articles.
@Max: It is just an investing style. Some people trade based on speculation. If they feel the stock is going to go up today, or maybe for a few days, they will buy and sell it quickly. It works when you have a lot of money, else wise transaction costs eat you up. Honestly day traders tend to lose in the tend no matter what
The insider trading charge is problematic. High frequency trading is done by computer, not humans. So the humans who wrote and/or deploy the code are the ones to profit. However, to prove insider trading, you have to prove that the *humans* had the information before everyone else. But they didn’t. Their machines did. And you don’t want to set a precedent that holds humans accountable for the data on their machines. If there were such a precedent, then you could be convicted of various hacking offenses…for being infected with a virus.
The whole thing reminds me of an old quote about corporations: Neither bodies to kick nor souls to damn.
And each of those transactions should be taxed, which would eliminate our national debt quickly even if the tax rate were minuscule.
XUUTHS – I agree. Even a 1c per transaction fee would help given the volume of trades each day. I do know that day trade gains should be taxed at normal income rates since they are considered short term gains vs the capital gains rate incurred on long term (1+ yrs) gains. I also know like Michael stated that most day traders are not much more then gamblers and except for a few lose more then they make.
XUUTHS — many countries have a small transaction fee on stock trades, and I think it is entirely appropriate. According to Wikipedia, there are 40 countries that have some kind of Financial Transaction Taxes (FTTs), which together have raised $38 billion. Britain has had one since 1694! The US does actually have one, but it is 0.0034% (very small).
Michael, you are dead wrong about insider trading. Insider trading is trading based on information that is not public. It matters not whether that trade is executed by a machine or a human. Let’s take a simple example. If I, as a human, know that Apple’s earnings are not going to meet expectations and I sell my stock, that is insider trading. If I program my computer to receive Apple’s earnings before they are public and program it to sell my stock if the earnings are below expectations, that is exactly the same thing. The next thing you are going to tell me is that if I program a computer to kill someone, then I’m not guilty of the murder.
IK, your examples are inapt in this case. Your quotes: “If I program my computer to receive Apple’s earnings before they are public…” “If I program a computer to kill…” Both of these examples involve intent on your part. You did something (actus reus) with the intent (mens rea) of causing death or fraud.
That’s not what happened here. The code was written with legitimate intent of using information released through legitimate channels. There is neither actus reus nor mens rea. The problem is that the information was released to the legitimate channels early. The traders cannot be held responsible for the actions or mistakes of someone else. If I am driving 70 MPH on the interstate and someone jumps off an overpass, landing 2 feet in front of my car, I cannot be convicted of murder, manslaughter, or anything. I was performing a legal action and had no intent to cause that person’s death.
If it were the case that the traders somehow caused the early release, that’s a different situation entirely. But no one is suggesting that is the case here. Unintentionally benefiting from a technical glitch is not a crime.
For the record, I am not defending the practice of high frequency trading. I think this case is a perfect illustration of how it creates a very volatile and dangerous marketplace. Anyone who has studied distributed systems can tell you that synchronous behavior in such environments is impossible to guarantee. But just because I think it’s a bad idea in general doesn’t mean that this particular case was a crime.
Ah, I see the root of our disagreement. You say “If it were the case that the traders somehow caused the early release, that’s a different situation entirely. But no one is suggesting that is the case here.”
Actually, there are people saying exactly that.
They are trying to figure out how this happened. Somebody must have programmed a computer to receive this information, and the only people who knew the information were in a quarantined room and were sworn to not release the information. It could be easily argued that programming a computer (before the release time) to receive that information IS a release of that information.
The bottom line for me is that the announcement is just words. Somebody had to translate that announcement into information receivable by a computer. Somebody must have done that before the release time in order for the computer to execute a trade in less than 3mS after the release time, which sure sounds like a violation of the release rules and might even be intent to commit fraud.
“Actually, there are people saying exactly that.” Yup, we’re just interpreting the articles differently, because I’m still not getting that.
My take is this: There are two groups involved, the traders and the news organizations. The news organizations are bound by the embargo, and it appears that someone within the news organization moved the data from New York to Chicago ahead of the 2:00 time specification. The news organization’s Chicago office then released the information at exactly 2:00 Chicago time. The traders’ code, then, parsed the news release (after 2:00) and initiated the trades immediately. Your view is that this can only happen if there is collusion between the news and traders. I’m not convinced that is the case.
While I am not familiar with the intricacies of low latency news releases, I find it plausible that they tend to follow a predictable structure. If so, I think it would be quite possible (given the news release is of small size) to parse it and perform a quick natural language processing algorithm on it, scanning for very specific keywords, and initiating a trade as a result. This is where you and I disagree. You find it implausible that traders are writing and executing such generic code without having advanced knowledge of the specific news release. I do find it plausible that people are doing this.
The reason I don’t see that there are accusations of collusion is from the following passages: “A key question is whether or not any [news] organization transmitted information out of the lockup room and into its own computer system before 2 p.m. … It is not clear whether that would violate the Fed’s rules.” The takeaway that I got from the articles is that they were concerned about the news organizations violating the rules, not that the trade was initiated so quickly. I didn’t get that there was an implication of collusion. It will be interesting to see how this plays out.
As a tangent, I think this raises the possibility of an interesting future lawsuit. If the Fed continues the practice of releasing information only from New York, a high frequency day trader with machines in, say, LA or Hong Kong could argue that the practice is unfairly biased in favor of New Yorkers. Thus, the combination of this practice with the legal protections barring insider trading turns out to create unequal protection under the law.
On another note. I find it funny these people had to wait until things were announced (or found out right before they were announced or whatever) and acted then. Anyone with any investing sense knew they werent going to touch QE