At this year’s TradeTech USA conference, held in New York on a pier on the Hudson River, gaming by high-frequency traders was the hot topic. Brokers were under the gun to defend their order handling and explain the steps they were taking to protect their customers’ orders from opportunistic HFTs. And, for the most part, that’s what they did.
But it soon became clear that thwarting the HFT was not just the responsibility of the broker and his technology. Brokers pointed out that it was up to the buyside trader to play his part, as well. Often, the actions of the buyside trader will work against successful order handling, brokers complained. In a nutshell, what the sellside wants is more discretion over the handling of the order. They need the ability to switch gears when necessary.
"You can put the same sophistication the high-frequency trader puts into his execution logic into yours," Owain Self, UBS’s head of algorithmic trading for the Americas and the EMEA region, said at TradeTech. "And you can be totally unseen in the marketplace."
But a well-designed algo requires flexibility to change course if necessary, Self explained. And that may not be possible if the institutional customer has his own agenda. Oftentimes, Self explained, a buyside will insist on being a certain percent of the volume, regardless of market conditions. Staying at 10 percent of a day’s volume, for instance, throughout the day, could amount to a signal for the tape watchers.
"We are being very predictable by being a certain percent of the volume," Self said. "We create a signal. They feed off that and trade more volume. We react to the increase in volume and create a higher signal. It’s a perpetual thing. We create more and more signals, and they make more and more money."
Self contends that it is essential to build some discretion into such a "participation rate" algorithm. If the client wants to be 10 percent of the volume, it is important to be able to reduce that amount to, say, 5 percent when being gamed, and increase it to, say, 15 percent when not.
Avoiding HFT gaming requires knowing when and where to trade, brokers say. And knowing when and where to trade requires following the signals the market is throwing out or that the broker has designed. But acting on those signals can be impossible if the restrictions the buyside places on the sellside are too onerous.
"We’ve been given a number of shares that need to be traded," said Amit Manwani, Nomura Securities’ head of quantitative and analytics products in the U.S. "The only way we can actually add value is if the client gives us some discretion. Then we can listen to those signals and provide liquidity when it is appropriate to do so. Or demand it when it is favorable to do so." The exec added that this is a key part of Nomura’s discussion with its clients.
Both the HFTs and the brokers listen to signals, but the broker is not able to react to them as fully as the HFT. That’s partly due to the disconnect that exists between the portfolio construction process and the execution process on the buyside, Manwani explained.
At the prop shop, the two are closely tied together, if not the same thing. At the money management firm, when the portfolio manager decides to buy, for instance, may not be the best time to conduct a trade. "There is no feedback loop," Manwani said, "that might pause the portfolio manager. The time scale over which the portfolio manager expects to earn his alpha is very different."
The sellside is further hamstrung because the buyside frequently doesn’t divulge the full size of its order. To avoid tipping its hand, the buyside will often just inform the sellside of part of its order. The buyside trader might even spread the full order around to different brokers. This lack of information makes it harder for the sellside to know how fast to trade the order, Self said. They are in the dark as to how aggressive they need to be.
If for example, the fill is falling behind, the buyside may override the controls of the algo or switch to a more aggressive algorithm. If the broker objects, the client could say, "I want to buy this now. You have no idea how much I have behind or what my objective is. I can’t describe that to you by the order parameters."
In the battle of wits that is trading, the central difference between the broker and the high-frequency prop shop is motivation. The HFT trades because he wants to. The broker trades because he has to. The HFT has full discretion. The broker does not. This puts the advantage in the HFT’s court.
"The high-frequency trader has unlimited discretion," Self said. "I don’t. That’s why he doesn’t get gamed anywhere near as much as I do."