How fast do you read? How fast can you flip through pages? How fast can you locate a specific number within those pages? How fast can you interpret the meaning of that number and execute a stock trade based on your interpretation? “Pretty fast,” you say?
Alright. I’ll grant you can perform all those activities with both speed and accuracy. But answer me this… can you do all those activities and place a stock trade at the end of them in less than one second? Highly doubtful, I’m sure you’d agree.
Yet evidence is regularly presented attempting to convince us that there are not a few but many traders who are doing precisely that… receiving corporate financial reports, sifting through them to find just the data they need, interpreting that data to determine whether to buy or sell shares, and finally placing a trade according to that interpretation – all in less than a second, sometimes even as quickly as within 50 milliseconds, or one twentieth of one second.
Naturally, much of the process is automated, minimizing human intervention and thereby speeding up the process. But there is another explanation that must not be overlooked – for where the first explanation (automation) is legal, this second explanation is illegal. It is thus vital we seriously consider this second explanation if we want to level the investment playing field.
Let’s first take a look at why the first explanation of automation doesn’t quite measure up.
The High-Speed Trading Edge
“Some subscribers to data feeds for SEC regulatory filing are able to trade on the information faster than other investors, research shows,” reported the Wall Street Journal yesterday, citing a suspicious trade in Balchem Corp. (NASDAQ: BCPC) on November 9th, 2012.
As depicted on the chart below, Balchem’s report was made available at the SEC website at 1:45:48 pm; that’s 1:45 and 48 seconds. However, a spike in trading of 3,500 shares occurred at 1:45:24, some 24 seconds before the company’s financial report was made public.
Source: Wall Street Journal
The explanation? Subscribers to direct feeds received the company’s report at 1:45:24, and had a 24-second advantage over the rest of the investing public.
But that explanation does not hold water. Why? Because the spike of 3,500 shares did not take place “during” those 24 seconds but right at the “start” of that lead-time – within the very first second, within the very first milliseconds of that second.
Had the spike occurred closer to 20 seconds after the direct feed, then it might be physically possible for someone to scroll through the report, find the right data, interpret it, and fire off a trade manually.
Had the spike occurred closer to 5 seconds after the direct feed, then it might be electronically possible for a computer program to sift through the digital report, find the right data, match it with pre-coded buy or sell instructions, and fire off a trade electronically.
But to have a computer program receive the feed and go through the whole process of searching, isolating, comparing, matching, and executing a trade within a few milliseconds?
Naturally, high speed computer programs are powered by some lightening-quick processors with ultra-high speed cable connections to stock exchanges. So theoretically yes, it is possible to run through that whole procedure in fractions of a second. Studies cited by WSJ confirmed it:
“The studies are the latest indication that some superfast, sophisticated trading firms are enjoying an advantage over other investors, echoing previous cases in which high-frequency traders received corporate news releases or key data on the U.S. economy milliseconds before competitors,” WSJ informs.
Other similar incidents of spikes in trading volumes occurring immediately upon the release of reports via data feeds have pin-pointed such trades to less than 50 milliseconds of the release of the reports, or less than one-twentieth of a second. As a comparison, an eye takes 300 milliseconds to blink.
Super fast processors can compute trillions of calculations in a second, that is a certainty. And light travels 200 miles in just one millisecond. So, yes, theoretically it is possible to receive a digital financial report, sift through it, isolate key data, match it with the right instruction, place a trade, have it travel to the exchange which can sometimes be only a few city blocks away, and have the trade executed – all in less than 50 milliseconds.
But if the Balchem trade on November 9th, 2012 was such a great opportunity, with thousands of data feed subscribers nationwide, why were only 3,500 shares traded? At some $32.00 per share, it was a mere $112,000 trade. Does that sound like an order that high-frequency trading computers scouting great opportunities would place? Or is it more along the size of order you might expect from a corporate insider?
What is Really Happening
Another soon-to-be published study cited by the WSJ examines 4,782 cases in which corporate insiders purchased stock of their own companies. Of, course, there is nothing illegal about corporate insiders purchasing shares in their own companies, as long as the trades are placed after the information is made available to the general public.
What is interesting about these cases covered in the study is that “in 57% of those cases, the subscriber received the information before it was publicly available.” That is to say, in 57% of those cases the insider trading their own company stock was also a subscriber to an early data feed.
Does anyone smell smoke? Might it be that these high-speed trading programs and the data feeds that supply them with information are just a cover for insider trading? Could insider traders simply be using the two systems as a veil to cover what is really going on… acting on insider information before it is made available to the general public?
Of course, one would argue that disseminating financial reports via data feed is making the information available to the public just as when the SEC publishes the reports online. The only difference is that the data feeds are faster than SEC-online. But in both cases the information is public, and subscribers to high speed data feeds should not be penalized for paying to access public information a few seconds faster than other investors who choose not to subscribe to the service.
But is that really what they are doing? Are they really acting on information available by data feed – which is public information – and then using it to execute a trade? Or have they instead already cued up a trade, knowing what they want to do a day or two earlier, and are simply waiting for the data feed to arrive to trigger their insider trade?
What I’m asking is… have they already committed an insider trading crime by having made a decision to buy or sell shares of their own company based on information they knew at the last board meeting, and are now simply waiting for the data feed to place their insider trade in the “safe” window of trading times?
If They Really Wanted to Level the Field
Unfortunately, this is a question that cannot be answered, because no one can read another person’s mind. We do not know if an insider who placed a trade within milliseconds of the release of data through an early feed had simply let their computer make the decision or if they had already made that decision a day or two prior based on insider information only they and their board had access to. We can prove when a trade was placed, but we can never prove when the decision to place the trade was made.
If we really want to level the playing field for everyone there is only one solution… connect the data feeds to the SEC website, so that as soon as the SEC releases the information, everyone gets it at the same time, whether they get it from the website where they have to manually scroll through it, or they get it digitally via data feed where their computers can zip through it much more quickly.
Of course, even that would not stop insider trading, since a board member or company executive would simply wait until the report appeared on the SEC website and then transmit their insider trade within milliseconds of that start time. They would still be beating the rest of the marketplace, since everyone else would still have to scroll through the report, isolate the data, interpret it, make a trading decision, and manually input a trade. The insiders would have the first few steps taken care of days ago, leaving them only the final step of actually placing the trade.
So even if we hook up the electronic data feeds to the SEC website, we still wouldn’t stop the insiders from having a jump on outsiders. What we would stop, however, are the complaints, the suspicions, and the rambling commentary like this one.
Joseph Cafariello
