05:05 PM
PDQ ATS's Pause & Algorithmic Trading Crowd Attains Volume Milestone
The dark pool operator PDQ Enterprises said that its venue PDQ ATS captured more than 1% of US equity trading volume in August through its internal matching and custom routing capability.
For the month of August, PDQ reported total volume of 1.13 billion shares, accounting for 1.03% of the total US equities volume of 109.9 billion shares that month. In daily numbers, PDQ executed around 53.8 million shares a day in a market where the average volume was 5.23 billion, said Keith Ross, the Chicago company's CEO.
"I don't think there was any particular big macro jump," he said. The venue has been steadily increasing its volume. In July, it was probably executing about 0.8% to 0.9% of the national volume.
It appears that PDQ's customized routing capabilitiy is responsible for most of the volume gains.
According to a Rosenblatt Securities Let There Be Light report (registration required), PDQ executed 4.5 million shares on average in July through its matching engine. This does not include volume executed through PDQ's customized routing capability.
Reaching this major milestone is significant for PDQ. It shows that traders continue to migrate to the venue, whose process simulates a high-speed, algorithmic version of a trading floor.
Ross attributes the increase in volume to the addition of a few clients on both sides of the market in combination with slow, steady growth. Participants include "liquidity providers that are typically high frequency and stat-arb types, and on the liquidity-seeking side, clients are day-trading arcades, agency brokers, and other dark pools that ping us for liquidity."
The steady growth of volume is also a sign that PDQ's model is resonating with traders who are concerned about gaming, front running, and other issues in today's high-speed electronic equity market structure. With questions in the market around wholesale brokers that match trades internally and how they are handling retail orders, PDQ is in a position to make wholesalers work harder. "As they compete, the best price will win."
Though retail brokers like Charles Schwab have said they are happy with sending their orders to wholesalers such as Citadel Execution Services, and with the markets they get, Ross said PDQ ATS has a dozen market markers competing with one another. "It's not just one player making a market. It's a competitive group."
To recreate a virtual trading floor, PDQ's process includes a brief "pause" to allow for liquidity aggregation through an electronic algorithmic trading crowd of automated market makers competing for order flow. "As soon as a client is willing to pause their order -- we have a 20-millisecond and 5-millisecond pause -- we can use that time to aggregate liquidity. We believe that our process is such that, during the pause, we can not only make the market makers compete, but we can aggregate that."
PDQ's pause: Market makers respond
Another dimension of this auction process is that, because of the pause, the liquidity provider is responding to a marketable order for a particular moment in time. For example, they are being asked, "For the next 18 milliseconds, what is your market in IBM or SPY?" Making a market in a very short interval under specific conditions is something the PDQ algorithm can respond to, and it changes the risk profile. Because entities are opting into the trade, they can make a market in a bigger size if they want to do so, Ross said.
This is different from posting in a lit market, where market makers will seed their book with orders to compete with the next trader. He offered an example: A market maker posts on Archipelago or the NYSE where Microsoft is 46 bid at 46.01 offered, and 24,000 shares are up. The frustration is that, though the first trader there may get 1,000 shares, the second trader may not get anything. "You will post much smaller size, and you will also start to cancel when the market has a chance to move adversely."
Comparing this to a game of musical chairs, Ross said it takes a human hundreds of milliseconds to hit the market, whereas the market maker can get out of the way faster.
All that falls away in the PDQ model, he said, where the market maker is asked to respond to a marketable order. Instead of having to populate the book at the different exchanges and vie for the queue position -- an important strategy for high-frequency traders in lit markets -- at PDQ, market makers don't post a market unless they know a marketable order is there. "And when market makers post a market, it's because their algorithms think it's a good time to make a market."
Now PDQ is looking to add more institutional clients to its roster of subscribers.
During the first quarter, PDQ hired two veterans of institutional trading, Jim Ross and Peter Jenkins, to market its trading venue to the buy side. "Our talking point with the institutions is that their order is what drives the market," said PDQ's CEO. "And if they understand the power they have in deciding where to go with their order, they can fundamentally change the market structure." He believes the venue's volume will continue to rise as a result of trader concerns with high-speed trading.
Ivy is Editor-at-Large for Advanced Trading and Wall Street & Technology. Ivy is responsible for writing in-depth feature articles, daily blogs and news articles with a focus on automated trading in the capital markets. As an industry expert, Ivy has reported on a myriad ... View Full Bio