HFT, Reverse Splits and Hidden Signals
Jay Saluzzi
Themis Trading LLC
10 Town Square, Suite 100
Chatham, NJ 07928
973-665-9600
www.ThemisTrading.com
>
There is sadness in the HFT world today as they are about to lose their poster boy, Citigroup. Yesterday, Citigroup announced that they will attempt to get their stock price higher and try a 1 for 10 reverse split. We don’t know if it will work but we do know that close to half a billion shares of meaningless volume is about to exit the market. The WSJ had this quote:
“It’s going to sting,” said Joseph Mazzella, managing director for equities trading at Knight Capital Group. For high-frequency traders, in particular, he said, “it’s going to have a big impact.”
We are not exactly feeling sorry for the HFT guys as we are sure they are cooking up some ways right now to offset this lost income.
Speaking of HFT, we found a very interesting recently published academic study titled “A Dysfunctional Role of High Frequency Trading in Electronic Markets“. Click here to read paper 
Obviously, the title grabbed our attention as we finally found an academic study which could not have been funded by an exchange or large brokerage firm since the conclusion was not that HFT was great since it shrinks spreads and increases liquidity. The authors, Jarrow and Protter, are two very distinguished Cornell professors that have a long history of research in the financial field (unlike the Brogaard study which was published by a “candidate” for a finance PhD). Here are some highlights of their report:
-High frequency traders can create a mispricing that they knowingly exploit to the disadvantage of ordinary investors.
-High frequency traders see a common signal that they then transact instantaneously on before the signal is incorporated into the market price.
-Since all HFT sees the same signal, they all do the same trades at the same time. They create their own momentum which generates profitable returns.
-High frequency traders’s trades cause the price movement creating a self-fulfilling profit.
-The authors liken this activity to market manipulation from large traders except that with HFT profits are unknowingly generated via a market signal.
The authors appear to be saying that when HFT sees a signal in the market, they all act immediately and simultaneously to trade off the signal. However, the authors don’t tell us what the signal is that HFT is using to extract there profits. They say the signal could be the difference between the futures and forward prices of a stock index but don’t say exactly for sure. We have some ideas on this that we will share with you in future posts.
~~~
Tweet
Facebook
Reddit
Digg this!
March 22nd, 2011 at 8:13 pm
So, this is the latest post from Themis trading to lobby against HFT (so as to protect Themis’ business). Unfortunately, the post from Themis contains a number of cognitive biases and just plain errors.
They quote a paper from “two very distinguished Cornell professors” (a basic cognitive bias of appeal to authority, as well as a basic error, one of the professors is at Columbia, not Cornell – sloppy work from Themis).
The post refers to a study, saying “we finally found an academic study” where “the conclusion was not that HFT was great since it shrinks spreads and increases liquidity.” I find it remarkable that Themis make plain that they have been searching for a study to support their own lobbying, and that they “finally” found one. This is called confirmation bias, but it is rare that lobbyists admit to it so readily. Or foolishly?
The Themis lobbyist then goes on to try to discredit all of the other studies, saying the overwhelming plethora of research studies that conclude the benefits of HFT are “funded by an exchange or large brokerage firm”. Really, such childish appeals to the desire to find conspiracy theories is tiresome. What do you expect from a lobbyist, though?
The Themis lobbyist then goes on to selectively quote tiny portions of the research. What the Themis lobbyist doesn’t quote, though, is this line in the paper: “The existing empirical literature, although mixed, supports the conclusion that it improves market efficiency by reducing bid/ask spreads and market volatility while making markets more liquid.”
Let’s read that again from the paper; HFT “improves market efficiency by reducing bid/ask spreads and market volatility while making markets more liquid.”
The paper quoted is a valuable one; it is just misused and misquoted by the Themis lobbyist. For example, in the paper the authors attribute some of the profitability of HFT to “the differential speed advantage of high frequency trading that causes the inequity. To the extent that the speed advantage is generated by preferential treatment in the execution of market orders, these can be eliminated.” This is a good point, preferential treatment should be removed if it helps to reduce the unevenness of the playing field.
I urge anyone wanting to look beyond the crude Themis lobbying efforts to read the whole paper, it can be downloaded here: http://papers.ssrn.com/sol3/Delivery.cfm/SSRN_ID1781124_code362528.pdf?abstractid=1781124&mirid=4
March 23rd, 2011 at 8:34 am
I am Sal Arnuk of Themis Trading. Joe and I write all our work, do our own research, and as a matter of fact, hit on the ludicrousness of “lobbyists” in our society quite often. So, I promise you we do not hire or use any lobbyists to make our arguments. We prefer to make our point using logic and passion, as opposed to using dollars to purchase representation for your point of view against other citizens.
We tire of the silly “shrinking spreads and increasing liquidity” argument. Liquidity does not equal volume, no matter how many different professors you buy off with grant money to state that. The silliest thing? How the exchanges have forsaken America for their meaningless short term bottom line. For-profit exchanges do not work. They do not serve the intended purpose of exchanges. Read our blog post this morning: http://blog.themistrading.com/?p=2266
March 23rd, 2011 at 12:53 pm
This paper is based on artificial assumptions far removed from reality, such as:
1) “we assume that seeing the same signal, all high frequency traders do the same trade at the same time.”
ALL traders, SAME trade, SAME time?? This is simply wrong. How about “SOME traders do the same trade at the same time.” More realistic, but less suitable for making far-reaching generalizations.
2) “we provide a model with no bid/ask spreads and perfect liquidity” (the authors assume in the paper that all HFT traders always take liquidity, at zero cost).
This is more than plain wrong. It’s totally bogus to make an analysis of market microstructure claiming to represent reality while assuming a market with zero bid/ask spread and zero commissions.
In fact, most of the profit derived by HFT traders comes from providing liquidity (getting paid the bid/ask spread), not from riding the momentum. The zero-spread assumption turns the way HFT traders make money upside down.
3) The authors conclude: “in unison the high frequency traders create their own momentum,
which generates profitable returns.”
Well, the only way to make money by intentionally creating slippage / momentum is by being among the first to get into the trade, and among the first to get out, before the market impact of the trade bounces back. The fastest of the HFT traders can win in this game, but not all of them simultaneously. It’s a zero-sum game, since any artificial short-term momentum / market impact will eventually revert, and those late to the game will lose.
Realistically, the faster HFT traders playing the momentum game may feed off the slower HFT traders playing the same momentum game, but that’s about it. Don’t forget about other HFT traders who play the mean-reversion game – these are the folks who will kick in when price gets too far out of whack.
Here’s what I don’t get: why is momentum game evil at the intraday time scale and not evil at the multi-day scale?
I would argue that inefficiencies (a.k.a. bubbles) created by momentum traders / trend followers on the multi-day, multi-month, multi-year time scale are much more disruptive and harmful than whatever inefficiencies are created by HFT algos on the multi-second time scale. Think NASDAQ ’2000 or Housing ’2008.
Finally, a psychological observation: when you are angry at a certain way to generate profits (be it momentum trading, or mean reversion trading, or HFT trading in all its forms), you are not seeing reality as it is.
March 27th, 2011 at 1:05 pm
Sal is right, the exchanges we have and the whole equity market structure are anything but market where price discovery is transparent and anyone can trade at the NBBO. This is why every month a new dark pool springs up.
And you HFT guys simply are not getting it: no one is against you making money, we are against you causing flash crashes with algorithms gone wild due to the lack of liquidity that has disappeared exactly BEACUSE OF YOU.