OF FLASH CRASHES, HIDDEN ACTORS, AND NASTY IMPLICATIONS…April 5, 2019
Remember the infamous May 6, 2010 flash crash that sent Proctor and Gamble's stock plummeting, losing and then regaining massive amounts of share value in a matter of mere seconds? I do, and I've tried to warn about the dangers of High Frequency Trading and "dark pools" for some time, and I'm not alone. Professional traders have done so as well, although my concerns have been rather different than theirs. My concerns have been about the fact that algorithmic trading done by computers at high frequency means that markets in a sense are not reflective of human market activity, simply because they are not reflective of the speeds at which humans can do things. During the May 6 2010 flash crash, Proctor and Gamble's stock plummeted, and then recovered, in mere seconds, faster than the infamous 1929 stock market crash, which took a day of trading, with short sellers helping to drive the nose dive. Granted, in 1929, the crash was more or less across the board, and in 2010 limited to one stock, but imagine if 2010 had been across the board. Of course, they tell us that safeguards are in place to prevent that from happening. But there's a nasty question that lurks in the background there as well: cyber systems are not "ultimately and finally" secure, so imagine if the safeguards themselves were hacked.
With all those thoughts in mind, consider this article shared by Mr. H.B.:
It's noteworthy that Zero Hedge titles this article "The Scapegoating Continues", although its reasons for doing so remain rather unclear. Is Zero Hedge trying to raise the same concerns as I am, that is, is the scapegoating referred to the fact that certain programmers are being sought to be brought to trial, while the wider philosophical issue of algorithmic trading itself goes unexamined? I don't know, but this is a rather intriguing comment at the very end of the article that makes me suspect that might be a possible interpretation, which statement also will be the occasion for today's high octane speculation.
The article is about the prosecution of Jitesh Thakkar for his part in allegedly participating in programming designed to "spoof" the system. Except, there's a massive fly in the prosecutorial ointment:
Jitesh Thakkar, who is being charged with conspiracy to manipulate markets, never placed any trades. Prosecutors are claiming that the conspiracy occurred between 2011 and April 2015, beginning with an email exchange with Sarao where he agreed to build the software. For those keeping score at home, this means that Thakkar's alleged misbehavior occurred after the 2010 crash (though, to be fair, there have been several similar 'flash crashes' since, including the Aug. 24, 2015 'flash crash' where the Dow dropped 1,100 points during the first five minutes of trading. (Italicized emphasis added, boldface emphasis in the original)
All of this, according to the US prosecutors, is designed to send a message, not about the idea of algorithmic trading itself, but those involved in programming the algorithms that do the trading:
By prosecuting Thakkar, prosecutors said they're hoping to send a message.
"We will also seek to find and hold accountable those who teach others how to spoof, who build the tools designed to spoof, or who otherwise aid and abet the wrongdoing," the US Department of Justice said in a statement announcing the criminal charges in January 2018 against eight individuals, including Thakkar. "The government wants to send a message that you have to be careful how you design these programs" for traders, Wayne State’s Henning said. If Thakkar is found not guilty, the message will be that "you have to be careful about how you design these programs, but you don’t really have much to worry about," he said.
Fortunately for Thakkar, the bar for a conviction is high. To convince the jury that Thakkar is guilty, prosecutors will need to prove that he knew he was participating in an illegal scheme to dig markets. (Emphasis in the original)
Ok, clear enough. However, in the final paragraph of this article, there is a comment that has my high octane speculation motor running in overdrive; it's this:
Meanwhile, HFT traders on Wall Street have been largely ignored by regulators as the scapegoated Sarao, who made just under $1 million on the day of the flash crash, a pittance in the grand scheme of things. Though a former programmer for Goldman Sachs was famously convicted, then vindicated, then convicted again for "stealing" code, including some of the bank's HFT strats. (Italicized emphasis added)
In other words, the article's conclusion does strongly suggest that the real issue for Zero Hedge is the same as for me, the issue of high frequency trading itself. Notably, Goldman Sachs hovers on the fringes of a conviction of a programmer, who was convicted for stealing some of the famous investment bank's code, which gave an indication of its high frequency trading strategies. The implications of that remark are, I suspect, enormous, for the remark implies that it would take but little modification to a bank's own codes to create market manipulations. Or to put that country simple, algorithmic trading that is "legal" can easily function as a means of market manipulation. And that, of course, takes the pricing mechanism one step further at a remove from genuine human evaluation and market activity. It's ok for a big investment brokerage to do it, but not ok for "the little guy" who spots a weakness in the system, and exploits it.
As one might suspect, this has my high octane speculation motor working in overdrive once again, for this means that both non-state actors (such as the programmers under investigation as outlined in the article itself), or state actors with known cyber-security and hacking departments - China and Russia for example, but face it, one might as well include all the rest: India, Germany, France, the United Kingdom &c. - could also be involved in exploring HFT vulnerabilities for the purposes of market manipulation and even economic warfare. To draw an uncomfortable analogy, it's like walking into a casino, and playing a game of craps with weighted dice, or playing slot machines whose computer random-number generators have been hacked to the point they're no longer so "random." And of course, all this makes me wonder as well if high frequency trading algorithms might lie at the heart of the strange phenomenon of all the "suicided" bankers.
See you on the flip side...