Trading is the ultimate competitive enterprise and all traders, whether they are scalping ticks, long-term trend followers, short-term traders or day traders, want to get their price. For commodity trading advisors (CTAs), being able to consistently execute at or near their price is key and, for some, can be the difference between success and failure.
Ensuring quality execution is a little more complex in a world of high-frequency trading. While electronic trading made life easier for CTAs and improved the overall liquidity of markets, it added an element of anonymity to market making and fewer players feel comfortable showing their size. This has led to an increased use of "iceberg" orders (orders that only show to the market a small fraction of the total order) and high-frequency trading algorithms that, at times, attempt to identify icebergs and exploit them.
"That is problematic if that is what you are relying on," says Michael Geismar, principal with Quantitative Investment Management (QIM). "Those are the ones — iceberg orders — that bots [algorithms] can detect and that is when you can get taken advantage of. If somebody knows that there is additional size behind that 10-lot order that keeps popping up, people keep moving away from that trade. If I place an iceberg order and it shows 10, but there is really 500 behind it, if I hit 10 and I see another 10 pop up, then I know or can guess that there is more behind that. Then people are going to stop hitting and they move away from you, and you are going to incur more slippage."
Bradford Paskewitz, founding principal of Paskewitz Asset Management, uses the iceberg concept, but has created his own proprietary method. "We have developed our own execution systems with a FIX engine and we use our own algorithms to split up orders," Paskewitz says. "We do break our orders up into smaller pieces and execute them algorithmically; it is an iceberg-like idea. You want to get it done in pieces and you don’t want to show size along the way."
That is easier said than done, as Paskewitz says the high-frequency traders (HFT) are looking for those orders. "The whole reason people don’t show size is because of those high-frequency traders," Paskewitz says. "You want to randomize your orders because if you keep putting out exactly the same order in smaller increments, it is easier to recognize the signature and people say ‘oh look, this same guy is doing all these orders’ and they will monitor you more; if you randomize your sizes and your times, it is harder to tell it is all one person."
While there are simple algorithms to break up orders, it is hard to stay ahead of the HFTs, and that has provided an opportunity for software firms to help managers develop execution algorithms that promise to reduce slippage.
"In every change in market structure, there is a beneficiary and there is no doubt that Portware and firms like ours are a beneficiary of the changes that are happening in the listed derivatives markets," says Scott DePetris, chief operating officer at Portware, a software firm that designs execution algorithms.
DePetris says he has seen the effect of HFT traders play out, first in equities and now in futures. "We are seeing eerily similar market structure behavior. Spreads are fairly tight in the more liquid listed derivatives space, but the visible liquidity has decreased, a lot of it is hidden like [iceberg] orders," DePetris says.
And it shows in his business, as 50% comes from futures compared to just 15% three years earlier, when most HFTs concentrated on equities and options.
"If you are showing a bid or an offer of size, the high-frequency strategies are digesting that and skewing the market, and you need to be cognizant of it. There are a lot of strategies that are looking for that information and moving the markets," DePetris adds.
Not everyone sees HFTs as a threat, though. "High-frequency traders benefit us," Geismar says. "We find ourselves, more often than not, on the other side of those trades, so we like to see more high-frequency traders. A lot of them are playing both sides of the market; they’re providing more liquidity to the marketplace."
Geismar says that because QIM’s pattern recognition approach is so unique, he doesn’t think their entries can be detected, but they still need to be careful because of their size. "For us, years ago we realized that we can’t execute as quickly as we would like. If we were managing $20 million instead of $5 billion, our execution style would be much different. There is inherent slippage with every additional dollar that you manage and with us at $5 billion now our most important goal with execution is to execute as quietly as possible. So we have to spread our trades out throughout the majority of the day."
Of course traders always have tried to mask their size, but the proliferation of HFTs has forced managers to vary their entries more often.
Paskewitz does not change his entry strategies all the time, but does look at factors besides size. "It is not like HFT where your half life is a few months. Some of the strategies haven’t been changed in years, but they get changed when and if they need it," he adds.
QIM has been successful in limiting their slippage despite their dramatic increase in money under management. "It is an ongoing battle, but we think over the years we have done a pretty good job of minimizing the amount of money we lose to slippage," Geismar says, while adding, "There is no real way to measure slippage because there is no way to know what the market would have done if you were not participating."
Paskewitz agrees that defining slippage is not a perfect science. "I look at the price right before we start trading and compare that to our average fill price and that is my definition of our slippage."
He adds that it is more of an issue, the shorter your time horizon. "You have to be more diligent the shorter-term you are trading, and in the spectrum we are shorter-term. I would say we have modest pressure compared to a high-frequency trader," he adds.
The more things change
Geismar is not the only one who does not view HFTs as a new and evil influence. Robert Moss, chairman and president of Ram Management Group, says, "The execution has become easier now because the markets are more transparent. The volume has to be placed on
the platforms."
While Moss acknowledges that pockets of illiquidity develop, he attributes it to traders not using limits and adds it isn’t anything new. "The markets can go into a vacuum. But that is not anything that wouldn’t happen on the trading floor either," he says.
Moss has the benefit of years of experience executing large sized orders on the trading floor. To him, HFTs are not some evil monolith attempting to ruin his day and make a living reverse engineering his program; they are simply traders doing electronically what traders always have tried to do, and that is to get an edge.
"There are programs on both sides of the market. It is a free market, and I think it is good. There is a lot of liquidity there. What we are trying to do is get a trade on in a certain area. I don’t find it to be very difficult at all," he says
Roger Show is principal of Livestock CTA, LLC and executes his livestock strategy through the same floor broker he has used for 25 years, but says that the onset of electronic trading has helped improve execution. "You have people arbing between the pit and electronic markets, so that adds liquidity to the markets. It is easier to move a 1,000-lot order than it was a few years ago. Even if you don’t use electronic trading, it still adds volume into the pit," he says.
For Show, HFTs are not as much of a nuisance as the long-only indexes. They can provide opportunity, you just need to prepare for them. "They have been around for a while, and you try and find out when they do it and try and take advantage of that situation when you can. You have to be aware of it," he says. "Everyone knows that if you are going to move 10,000 from the front month to the next, everyone is going to front-run you. There are ways to take advantage of the ETFs as well."
The positive thing about the long-only indexes is that they are transparent; you know roughly when they have to roll their positions.
That obviously is not the case with HFTs, but a trained eye can help. While some complain that no one shows size, Moss doesn’t see that as a problem. "It is there. The only thing you have to do is put an order in and buy it (see "A bigger pie"). My question is what is the [HFT] program trying to accomplish and how is it accomplishing it? If they are trying to sniff [stops out], how do you take advantage of it?"
The long and short of it
While the need to reduce slippage is a major problem for short-term traders, even the long-term guys need systems to help them execute. "They need, to the extent possible, to be as anonymous and as intelligent as they possibly can in entering and exiting large positions. You can see your entire alpha expectation degrade in a bad execution. They need to be focusing on that," DePetris says.
"As a long-term trend-follower who holds positions for months, it is not as critical," says Scott Hoffman, principal of Red Rock Capital Management. "My program’s performance still would be good with slippage [that] a short-term trader could not tolerate. With that being said, getting the best price for execution always helps," Hoffman says.
He points out that, as a long-term trader, he executes roughly 1,000 round turns per million managed and is making more than $100 per trade. "When I look at a short-term system, your average [profit per] trade is $20 and a tick in the E-mini S&P is $12.50; one tick slippage is half of your gross average trade," Hoffman says. "Our average trade can be a couple hundred bucks, so the same $10, $15 or $20 for slippage and commission represents a much smaller percentage of what we are trying to extract from the market."
It is a simple matter of math. "We all are trying to get gross profit/loss out of the market; how much do we have to pay in commission and slippage to get it?" Hoffman says.
This is critical for short-term traders like Revolution Capital Management. "If you look at the advantage that you have without any execution costs, you might have a better advantage with a short-term system," says Michael Mundt, a principal with Revolution. "If you look at any long-term system and you just add up how many round turns they are doing — assuming 2,000 round turns — they are going to pay approximately $20 a round turn, that would be $40,000 per year in execution cost per million. If you look at someone doing 10,000 round turns per million per year, if you don’t do it well that is 20% per year. So you just erased probably all of your advantage in trading costs," Mundt says (see "It is all in the numbers," below).
Revolution had developed a successful long-term trend-following system prior to launching their short-term system. The latter took much longer to develop, which Mundt attributes partly to the difficulty of execution.
Revolution’s short-term Mosaic program is trading four times as frequently as most longer-term managers. "With us, the potential downside is huge if we don’t make the trading efficient; it is simple arithmetic," Mundt says.
That is true for traders who trade more frequently and for those who have large size to execute. "We try and place as many passive orders as we can, and try and get filled in a passive manner. If we are buying, we are not [lifting] the offers because that can give you away," Geismar says.
And that seems to be key. Execute as quietly as possible and the size will be there, but no one is going to show the size so you have to exercise some patience. "It does seem like a constant battle to try and make sure that we hide what we are doing, randomize what we are doing and make sure we don’t send out signals to what we are doing inadvertently," Mundt says. "There is a trade-off that has to be made. Ideally you execute as quickly as you can, but if you do that with big orders, it is self defeating. There has to be some balance between speed of execution and minimizing your footprint in the market."
It may be difficult, but it is good to remember the basics, and what HFTs are trying to do is nothing new. "For every one side of the market that is doing that, there is another program that is doing that on the opposite side," Moss says. "These programs become effective initially, but over the long-run they lose their productivity and profitability. After a while there is the next new trading program that is taking advantage of what they are doing, if they are giving up too much of an edge."
