Interview with Sal Arnuk, Themis Trading

In the second in our series of High Frequency Trading Interviews, we talk to Sal Arnuk, Co-Founder, Partner and Co-Head of Equity Trading at Themis Trading LLC, an agency trading firm specializing in US Equities for institutional investors. Prior to founding Themis, Sal ran an agency stock trading business at Smith Jacobs and Co, and before that he spent ten years with Instinet. He holds an MBA in Finance from NYU’s Stern School of Business and a Bachelor’s Degree in Finance from SUNY Binghamton University. A regular commentator on the subject of High Frequency Trading, Sal recently attended the SEC’s Market Structure Roundtable.

High Frequency Trading Review: Sal, how would you define high frequency trading?

Sal Arnuk: I actually don’t think it’s useful to lump everything into one bucket called high frequency trading. Everyone does it just because it makes for an easy categorization, but I guess you can call HFT any strategy that involves high speed execution of thousands of orders per second, or in a short time period. I don’t even want to handicap it by calling it thousands as some people can do a million.

But the point is, some of this activity that people call high frequency trading is positive and beneficial for the markets. I’ll give you a great example, statistical arbitrage. Where for example IBM is trading at one price in London and it’s trading at a different price on the New York Stock Exchange, a slightly different price on another exchange and its option is trading at a slightly different “out of whack” price level in the options market. So a statistical arbitrageur will capture that inefficiency and sell the market that is more expensive and buy the market that is cheaper, in the process earning a small profit for himself while tightening spreads and increasing liquidity, which are good things for nearly all investors.

HFTR: So you would see that particular aspect of high frequency trading beneficial, because it is increasing liquidity and removing inefficiencies in the market?

SA: In a non-predatory way. That is the key. What I described is an example of high frequency trading which is valuable and serves a purpose.

But there are predatory types of high frequency trading and frequently the terms used to describe the predatory aspects are “momentum ignition”, “footstep detection”, where shops have co-located advantages, i.e. high speed advantages, by renting space from the exchange so the exchanges are complicit in profiting, it’s part of their business model. They sell them the speed and they sell them the data feeds that make it very easy to infer institutional footsteps. When I say institutional, I mean pension managers, mutual funds, large hedge funds, retail investors.

HFTR: So do you mean that high frequency traders using these predatory types of practices get to see orders before the rest of the market and are able to take advantage of those institutional orders?

SA: No, in this data feed issue that I am talking about, they are able to tie back. Let me give you an example. Imagine you are a seller of Abbot Labs, a NYSE Stock (symbol ABT) and it’s trading roughly $47.60. Let’s say you are a seller at $47.60 for a very large UK pension fund. Let’s say you are the trader on that desk and you have two million shares to sell and you are selling stock anonymously in a machine at $47.60, two hundred shares, three hundred shares, a thousand shares, piecing it out.

Now imagine if a computer program is set up and has a written code using special data feeds from the exchanges that allows that program to actually infer and tie back each execution and figure out that someone over the last 18 minutes has sold 180,000 shares of Abbot Labs as low as 47.60 and even higher than that, but they wont go lower than 47.60. Would you agree that this is useful information? If you are a trader and if you knew there was a large seller, you would probably try to short stock ahead of that seller and then maybe buy it back cheaper. So this is useful information and this is the type of strategy that is just one type of predatory activity that we’ve seen in the marketplace. There are plenty of others.

HFTR: OK but the supporters of high frequency trading would say, well this is no different to what’s always happened on stock markets where you have specialists and dealers on the exchange floor for example, who see those big orders coming in and are able to infer what’s going on from the way the order is being executed. How is what’s happening now with the use of computers different to that, other than that things are faster?

SA: We would argue that it’s different in two meaningful ways. First of all, the exchanges aren’t complicit. The exchanges have changed their business model so that they don’t receive their primary compensation from listing companies. They now receive the majority of their revenues and their income and their profits from co-location, from data feeds, from market data revenue, from all these other activities that are very much catered and tied to high frequency traders. That’s the majority of their businesses nowadays. So why is it different? Because they are being supplied unfair advantages that did not exist in the past.

HFTR: That’s the co-location advantages and the raw data feeds?

SA: There’s the co-location advantages, number one and there’s also the data feeds, which is something that has been going on for the last several years but the lights have been shining on it only recently.

But It’s meaningful in a second way too. Let me ask you this, if it is wrong for a specialist to unfairly bastardize some order flow a certain percentage of the time, does it make it more right to do it in an amped up, hyper, across-the-board way? Because we have computers, does that make it right? If there is a practice that was wrong in the past and inefficient and should have been changed and certainly something that most of the buy side institutions when they talk about the specialist system would say “Jeez this was, it’s horrible when they do this and I hate it and it’s wrong,” and now it is being done in an exponential way, is that supposed to make it right? I find this to be curious logic.

HFTR: So what ideally would you have the regulators or the exchanges do? Would you like to outlaw co-location or would you like raw data feeds to have delays introduced so that they are in line with the SIP pricing distribution for example?

SA: Well this is another argument we’ve made, that there are different speeds between the public sector and the private co-located speed.

No, in terms of what we feel regulators should do, first of all we think that they should tag all of the high speed traders in the market. Just put a tag on them so they can study them and understand them, so that we do not have to go and listen to a “he said, she said” between myself and for example Mr. Narang from TradeWorx about who is right and who is wrong. Let’s all open our books and have the regulators look at it fairly and then we will see what’s wrong and who’s telling the truth and who’s not. So number one I think before anything should happen and any regulations passed, they need to be studied more because I don’t know all the answers and on the other side they don’t know all the answers either.

After that is done, I would also suggest that if we start just going about trying to solve the problem by legislating and regulating each symptom, it’s an inefficient way, it’s an inefficient process that is bound to have unintended consequences.

I would much prefer the root of the problem be addressed, which is that there are so many ample conflicts of interest in our markets today that were not there just five years, seven years, eight years ago and these conflicts of interest are the root of all these layers of malfeasance. Broker dealers and high frequency trading firms own each other, they own significant stakes in the exchanges and alternative trading systems. So should anyone really be surprised that broker dealers and high frequency trading firms that own stakes in the exchanges are getting special benefits from the exchanges at the expense of other investors? I’m not surprised. Is this not a red flag?

So this is what I’d suggest, number one, tag the traders so that you can study them more because we don’t want unintended consequences. Number two, when you do go about trying to decide what to do, rather than negotiate each band-aid or put a band-aid on all the leaking pipes, let’s get to the source of the leak.

HFTR: And the source of the leak is what, do you think?

SA: The source of the leak is once you get rid of this conflict of interest, between exchanges and the folks who own them, then you can have an appropriate regulatory function be performed at the exchanges. They can self regulate. Right now we’ve seen in the United States, be it with flash orders, be it with actionable IOIs, be it with the data feed issue we highlighted in our latest paper, which caused significant change in the UK and in the United States, they changed their feeds because of our paper. What will happen is if exchanges are not at conflict with some of their clients at the expense of others, then they can be relied upon to treat everyone fairly. But now every time their regulatory duty to provide a fair and orderly and equal marketplace runs back against their profit motives, guess who wins?

HFTR: So are you saying that all you would like to see at the moment is a proper identification of the high frequency trading activities and you are not calling for any more regulation than that at this stage?

SA: I would say as a start, yes. As a second point, I would want to get a handle on how highly the high frequency traders are leveraged. Because from a regulatory standpoint, they want to make sure there’s not a systemic risk. That there is a risk of a few of these firms blowing up and bringing down the market as collateral damage.

HFTR: That is an interesting point because my understanding is that most of the time the high frequency traders’ positions would be flat, other than for the very short-term periods they are in the market.

SA: Here is the thing, they start the day flat, they go home flat, they trade hundreds and thousands of trades, millions of trades each day. Hundreds of millions of shares, making very small margins. As more and more high frequency firms come into the space, what do you think happens to their margins? They shrink, each one is making less. They are sharing the same pie; they are eating each other’s lunch. So as what happens and by the way each of these firms have convinced their clearing agents that “look my activity is so risk free because I’m going home flat everyday, I am not taking risk, so give me higher than two times leverage, give me higher than four times leverage, give me ten times leverage”, so that one of these firms that is managing £50,000,000 is trading like £500,000,000. With that leverage, what happens when returns shrink on Wall Street, what do we know from history? They amp up the leverage. And when they amp up the leverage, it’s only a matter of time before one of these firms says, “this is a really hard business, let me take a small risk, let me go home long for one day. It won’t be a big deal”.

All it takes is for one of these guys to push the envelope, which our industry is famous for doing, and highly leveraged up, it has a systemic risk attached. So that’s why I think:

(a) tag the traders
(b) make sure we have their risk, their leverage under control and
(c) I would hold off on other regulations outside of circuit breakers until you have a better understanding of which activity is beneficial - because some high frequency activity is beneficial - and which activity is not.

HFTR: You mentioned circuit breakers there. Obviously there are a lot of investigations going on regarding the so-called “flash crash” of May the 6th and one of the things coming out of that is a call for wider use of circuit breakers. What is your view on what some of the exchanges did in busting open certain trades, and leaving some traders who expected to be flat, inadvertently long or short? Did the exchanges do the right thing there or should they have let all trades stand?

SA: No. First of all there’s the fact that they would pick such an arbitrary number to break trades, 60%. So basically if you sold the stock down 50% of its value, there is nothing wrong, but at 60% there is. The level they chose is so arbitrary and silly that that right there tells you it wasn’t a good idea. To tell you the truth, I actually feel - and Themis feels - that a trade is a trade and the public needs to have confidence in the markets. So to the extent that, think about the people who were disadvantaged because they were inadvertedly made short when they thought that they were covered, at a loss or a profit. It creates all kinds of other problems and bad precedents. A trade must be a trade.

But on top of that, it raises another issue. And the other issue is regarding circuit breakers, the clearly erroneous policies by all the exchanges are a web of non-uniform and arbitrary and a poor attempt at restoring fairness to erroneous trades and that had to be addressed. I think the circuit breakers are a good band-aid, they are a good start but they have to be uniform across all exchanges, all ATSs, all dark pools, all light pools, everybody without exception, hands down. If you don’t do that, you are just going to create an arbitrage situation where if trading slows down on one exchange, all the traders just rush to pile into positions on the other exchangers and that is going to cause a cascading effect and cascading volatility which will exacerbate the very situation we are all trying to rectify. The challenge with circuit breakers is if you study OPEC and the activity of group thinkers and group policies, it’s very difficult to get everyone to agree and act in a voluntary manner on one set of procedures. So for example NASDAQ came out immediately with it’s own suggestion, BATS came out with a different set of suggestions for the circuit breakers and what’s really needed is the “mama bear & papa bear” regulator to come in and just hand down the law to everybody and everyone just has to adhere to it.

HFTR: Any final thoughts?

SA: Yes. Everybody talks about the benefits of high frequency trading being lower trading costs, cheaper commissions, cheaper spreads, narrow spreads and we think that that is humorous because it is in the most liquid stocks that don’t need it and it’s only there when it’s profitable or guaranteed or close to being guaranteed to be profitable. Now is that really providing liquidity? We are so used to these two words, providing liquidity, liquidity provision, “oh we provide liquidity, we should be incented to provide liquidity.” When you provide liquidity, only when you make a profit and you pull your orders when there is market stress, that’s not providing liquidity. It’s scalping when times are good and running for the hills when times are bad. So I don’t mind it, I wouldn’t want to stand in front of a freight train if it were my money, so I understand it. But please stop the disingenious talking out of both sides of your mouth. You can’t have it both ways.

HFTR: Thank you Sal