There is one thing Jos Schmitt wants to make clear: Not all high-frequency trading is bad. During an interview in his Toronto office, Schmitt drops this caveat nearly every time he lambastes certain HFT practices. “There are strategies that HFTs can deploy that are predatory in nature,” he says. “That is the big issue, and that is what we have to remove from the market.”
Schmitt is the co-founder and CEO of Aequitas Innovations, which has the backing of some of the country’s biggest financial institutions, including the Royal Bank of Canada, CI Financial and OMERS. This month Aequitas is launching a new trading platform for Canadian stocks, called Neo, to thwart HFT operators, which use algorithms to exploit split-second price differences. It’s a concept another Canadian, Brad Katsuyama, has already championed on Wall Street, and one embraced by an emerging faction of securities markets participants beleaguered by the ongoing arms race for ever-faster trading technology. By using a “speed bump” to delay trades and a host of other tactics, Schmitt and his ilk hope to level the field for all participants.
Aequitas has other ambitions, too. It aims to launch an exchange for private companies and start a listings business, competing head-to-head with the dominant stock exchange operator, TMX Group, on just about every front. Its backers are financial markets insiders who have become frustrated by the state of affairs today and want to see more competition in securities trading. Aequitas and its proponents say they are shining a light on how mainstream exchanges have enabled the rise of an unfair marketplace. “They put their own profitability ahead of the public service they are providing to the economy,” Schmitt claims. “That, to me, is where it went wrong.”
TMX, of course, is not sitting idly by. “Aequitas clearly is looking to have a high-impact campaign,” says Kevan Cowan, president of TMX Markets and group head of equities. “When you’re the new entrant, you have the ability to say what you wish in order to try to create the headlines.”
For Schmitt, a fair-haired Dutchman, getting Aequitas off the ground is almost like history repeating itself. In 2008, he launched an alternative trading platform to the TSX, called Alpha. It came to capture about 20% of the volume in Canada. Schmitt was trying to set up a listings business when Alpha was swallowed by TMX in 2012. In a way, Aequitas is a chance to finish what he started.
Regardless of the politicking and petty feuds, the success of Aequitas will likely hinge on a fundamental choice every market participant will have to make: How is a stock market, that hallmark of capitalism, best served—with unbridled competition or mutually agreed restraint?
One day in early 2009, Aldo Sunseri wanted to buy a large block of shares in Bombardier. That was nothing unusual for Sunseri, chief equity trader at CI Financial. But when he punched the order into his trading terminal, he received only a fraction of the stock he requested. It was not an isolated occurrence either. His fill rates—the percentage of an order that gets completed—were declining. Eventually, he came to suspect that someone was anticipating his trades and snapping up the stock before he could, a practice known as front-running. He just had no idea how. “It was getting very frustrating for me,” Sunseri says.
He wasn’t alone. Other institutional traders across North America were finding themselves unable to fulfil entire orders, or prices would instantly move, causing them to buy shares at a higher price. Although CI couldn’t figure out what has happening, it did a few things to cope, none of which were ideal. “We used to always target a certain percentage of the average daily volume [of equities] to not have a big market impact,” says Eric Bushell, CI’s chief investment officer. “All of a sudden, we had to cut that in half. And it just kept shrinking.”
For about a year, CI pressed on without a clear idea of what was transpiring. Then Brad Katsuyama showed up in the firm’s Toronto office. Katsuyama worked for RBC Capital Markets in New York. He was in town to pitch a product he had developed, called Thor. It was a smart-order router, which figures out where and when to send stock orders to achieve the best price and execution. Over the previous year, Katsuyama and his team had wrestled with the same problems vexing Sunseri and Bushell. Katsuyama (as recounted in Michael Lewis’s Flash Boys) eventually cracked it. And in that meeting, he explained everything.
Fast electronic trading was nothing new. Traders on the floor of a stock exchange, waving slips of paper in the air and yelling like lunatics, were already being phased out in the 1980s, as firms adopted more efficient electronic trading platforms. Complex algorithms that analyze and act on market data faster than any human ever could gradually came to dominate trading activity. By 2009, high-frequency trading accounted for 61% of U.S. equities volume. What helped fuel the rise of certain HFT trading strategies was fragmentation in the marketplace. In the U.S., there are a dozen different trading venues and even more “dark pools,” marketplaces where the size and price of orders are not revealed to any of the participants. Some HFTs learned to exploit the time it took for orders to reach the various exchanges to get a picture of the market a split second before anyone else. They could see, for example, that one player was making a large order of stock at, say, $20 per share. They could then race to purchase shares in a different marketplace and sell it back to that investor at $20.01. To do this, traders needed an incredible speed advantage—these trades happen in microseconds—and couldn’t do so without some help from the companies that operate stock exchanges.
Because HFTs trade rapidly and constantly, they generate a lot of volume—and revenue—for exchanges, which began to cater to them. HFT shops paid to place their servers next to the exchange’s (in the very same room, in fact) to limit the distance their orders travelled, maximizing speed. At CI’s office, Katsuyama displayed a diagram of various exchanges and the routes stock orders took to reach them. “That was the penny-dropping moment for us,” Bushell says. “Then we were pissed.”
Thor promised to solve the problem by slowing down orders so they reached exchanges at the exact same time, preventing front-running. CI started using Thor and saw fill rates increase again. It’s not effective against every potentially harmful HFT strategy, though.
The forces that led to these HFT practices in the U.S. were at work in Canada as well. The TMX had a “co-location” business to allow traders to put their machines as close as possible to the exchange’s, and the market had become more fragmented as a handful of alternative trading venues popped up to compete with the TMX. Some of those markets sought out HFTs to generate volume. The biggest was called Alpha, and it was run by Jos Schmitt.
Strange events in the markets were also perturbing Greg Mills, the head of equities at RBC Capital Markets. Everything started to crystallize when Katsuyama phoned him on a Monday morning to describe some early work he and his team had done on Thor over the weekend. “When someone tells you something that’s incredibly technically complex and you’ve never heard of it before, it’s like, ‘Really? What are you talking about?’” Mills says. It wasn’t until he saw it mapped out on a whiteboard that he realized how simple and ingenious a solution Thor provided.
But aside from high-frequency trading, there was another event that led RBC to spearhead Aequitas. In 2011, the London Stock Exchange (LSE) made a bid for the TMX Group, which operated the Toronto Stock Exchange and TSX Venture Exchange. The looming acquisition set off a nationalist firestorm in Canada. A consortium of Canadian financial institutions banded together, branded themselves the Maple Group and made a counter-offer. Toronto-Dominion Bank, The Bank of Nova Scotia, Canadian Imperial Bank of Commerce and National Bank Financial led the effort, along with five pension funds. RBC was left out; it was an adviser to the LSE on the acquisition. The LSE and TMX realized they wouldn’t be able to garner enough shareholder support, and the acquisition died. The Maple Group eventually triumphed a year later.
RBC was then invited to join the Maple Group. “We had some extensive meetings with them, and we looked at their business model and what they were going to attempt to do, and it’s not something I believed we should be a part of,” Mills says. During a regulatory comment period, he wrote a submission to the Ontario Securities Commission, raising the concern that the acquisition would remove competition from the marketplace. Maple not only ended up with TMX but also snapped up the Canadian Depository for Securities (a clearing house) and Alpha—the only substantial competitor to TMX when it came to trading. RBC couldn’t suddenly change its stance. “You’d look like a bit of a hypocrite,” Mills says. Besides, Mills had genuine problems with the transaction. “It was largely being put together for the benefit of its own shareholders and really didn’t bring any solutions to the marketplace,” he says.
Mills told Doug McGregor, head of RBC’s capital markets division, and then-CEO Gord Nixon that he recommended turning down Maple. Both agreed. But they wanted to know what the bank could do instead. Mills didn’t know, but he felt the organization was uniquely positioned to figure something out. It was one of the original backers of Alpha and, partly due to Katsuyama’s work, was well versed in HFT strategies and had developed a market structure research group to probe issues in the market. (Katsuyama eventually left RBC to start IEX Group, which runs an anti-HFT dark pool and has ambitions to become a full stock exchange.) Mills started thinking about how to design an entirely new stock exchange for Canada. One of his early calls was to Jos Schmitt.
After spending his much of his career working at exchanges in Europe, Schmitt came to Canada in 2003. A few years later, all six of the country’s big banks, along with Desjardins, the Canada Pension Plan and Canaccord, joined together to launch Alpha, in hopes of lowering trading costs and bringing more competition to the market. They chose Schmitt to get it off the ground. He became a leading advocate for competition in Canada. But then TMX acquired Alpha. You might expect Schmitt, who’d tried to position Alpha as a viable alternative, to take mild umbrage at this. But Schmitt says he was optimistic at the time about what TMX could build.
And what of his connections to HFT? (This is something Kevan Cowan at TMX points out in an interview. “The Toronto Stock Exchange is about 15% HFT,” he says. “The founder of Aequitas, his last venture was over 40% HFT.”) Schmitt doesn’t deny his past. “I remember in the very early days of Alpha, when we were struggling a little bit, we looked at how we could involve more high-frequency traders because they were very good at igniting liquidity,” he says. At the time, there was little discussion about HFT’s effect on the market. That debate developed over a few years, he says, and when it started to heat up, his attention was elsewhere—trying to launch Alpha’s listings business and then dealing with the TMX acquisition. “We never really sat back to zoom in on the implications of HFT,” he says.
Unsurprisingly, for someone who had positioned himself as an adversary of TMX, Schmitt didn’t stay long after the acquisition was completed in 2012. When Mills got in touch late that year to discuss the idea of starting a new exchange, Schmitt wasn’t interested. He’d already done that. Mills suggested he step back, reflect on what he saw as the problems in the market and come up with potential solutions. Schmitt, a cerebral kind of guy, was intrigued. He hired a small team and got to work.
Schmitt produced a white paper that raised concerns about high-frequency trading, the lack of competition in the market and the “disappearance” of market makers, the banks and broker-dealers who traditionally stand by as passive buyers and sellers of stock to ensure other investors have a counterparty. Their presence helps ensure stability and liquidity, especially when things get hairy. But HFTs were making it uneconomical for market makers, so they were retreating, according to Schmitt. HFTs assumed the role of liquidity providers instead—but without any real obligations, leaving markets vulnerable, he argues.
Schmitt singled out other issues, too, such as the cost of accessing market data and the view that companies were going public too early. All of these problems could ultimately be traced to the for-profit model of exchanges, which used to be mutualized. “When they became for-profit organizations, they put their own profit ahead of what’s in the interest of the investor and the issuer,” he says.
Schmitt submitted his final report to RBC just days before he was to get married and leave for his honeymoon. His thinking turned out to be very much in line with that of Mills, who, upon Schmitt’s return, asked him to head up a new exchange. Acknowledging there were new problems to tackle, Schmitt agreed. Mills, who serves as Aequitas’s chairman, and a colleague went about roping in partners and found plenty of willing participants outside the Maple Group: CI Financial Corp., Barclays Canada, IGM Financial, PSP Investments, OMERS, ITG Canada and even BCE. Getting the governance structure right was crucial; no one firm owns more than 15% of Aequitas, and the hope of generating huge returns was not their primary motivation for investing. Instead, it was to address the issues Mills and Schmitt identified. (The total startup costs for Aequitas are between $50 million and $100 million.)
Schmitt’s first notion to tackle the HFT problem was to exclude HFT operators. They would be allowed to make liquidity, meaning they could stand by with orders to buy or sell stock and only execute when other traders were ready, but they could not take liquidity or trade when they wanted to. Aequitas presented this concept in its proposal to the Ontario Securities Commission in 2013, but the OSC rejected the idea. Excluding any class of market participant from an exchange contravened the principle of fair and equal access, it contended.
“You can imagine how bizarre that felt,” Schmitt says. “The reason we want to do this is to achieve more fairness in the market.” His reaction at the time was more severe. “We came very close to shutting down the project,” he lets on. He thought that if Aequitas couldn’t limit HFTs in some way, there was no point in continuing.
During a later meeting with the OSC, one of the commission staff asked Schmitt if there was a way to “disincentivize” HFTs. Schmitt left with that word bouncing around in his head, and he started thinking about how to charge HFTs a higher fee to trade. That proved to be unworkable, too, because the necessary fee would be egregiously steep. The team then toyed with the idea of combining fees with a speed bump to slow down HFTs. They mapped a 50-kilometre radius around the Greater Toronto Area, where all the country’s exchange servers are located, to determine how long it takes for stock orders to travel between them. They determined that a five-millisecond delay, combined with a fee, was enough to put HFTs on equal footing with other investors. The final twist came with the realization that HFTs could figure out a way around a delay if they knew how long it was, so the team decided to randomize it between five and nine milliseconds, about one-fiftieth of the time it takes you to blink your eyes.
Aequitas devised the workaround in about a week and submitted a new proposal to the OSC. The commission gave the company the green light last November after a comment period. Schmitt believes the reworked solution is actually better than the first one. “We transformed a blunt instrument into one that is much more effective,” he says. Schmitt found the answer to bringing market makers back through his experience with derivatives exchanges. Market makers on those exchanges are guaranteed a certain percentage of volume as a form of compensation. Aequitas will introduce similar incentives.
With the trading platform making its debut on March 27, the only question now is whether it will work.
Aequitas isn’t the first trading platform of its kind. Katsuyama left RBC to found IEX Group, which started operating in the U.S. in October 2013. Like Aequitas, IEX applies a delay (350 microseconds) to all orders so that HFTs are prevented from trading on information ahead of other investors. Today, IEX is operated as a dark pool. Last September, it raised US$75 million and announced plans to convert into a full-fledged stock exchange. Its track record so far shows there is demand for a new kind of trading platform; IEX’s share of traded U.S. equities hit 1% for the first time in November.
Like IEX, Aequitas is a lean operation.About 40 people work at its Toronto headquarters. Even when the company is running at full speed, Schmitt plans to hire at most 10 more. “We’re working at a fraction of the cost that the Toronto Stock Exchange would to have the same kinds of services and reliability that we have,” Schmitt says. He’s fond of taking shots at TMX Group when he can, but the pressure is on him to prove he can deliver when the company’s Neo trading platform launches this month.
The ideas proposed by Aequitas enjoy a lot of support, if the submissions to the OSC during the comment period are any indication. Fund manager Eric Sprott backed the proposal while denouncing HFT. “We don’t have markets. We have manipulation,” he wrote. “I know in my gut that HFT is the death knell of markets.” Scotia Capital, whose parent company is a member of the Maple Group, did raise concerns with some aspects of Aequitas’s proposal but also wrote about its disappointment that “recent innovation by Canadian market operators had mainly taken the form of changes to trading fee models rather than attempts to improve the trading experience.”
Probably the most controversial aspect of Aequitas’s approach is how it deals with high-frequency traders. With the speed bump and additional fees, you might think HFTs would simply avoid the platform. But in Canada, regulators instituted something called the “order protection rule” a while back, which mandates that brokers direct trades to the markets displaying the most advantageous price. (It’s designed to protect investors from unscrupulous brokers.) Because of the rule, traders running certain high-frequency strategies could be forced to execute trades through Neo and pay fees required of no other investor, in addition to facing a delay. In its submission, Scotia Capital called the scenario “unreasonable.” Jenny Drake, principal of the quantitative investment team at Connor, Clark & Lunn, wrote that the approach amounts to Aequitas’s picking winners and losers. “We think it may actually be worse for the market,” she wrote.
Schmitt argues these criticisms miss the point. “The reason we need to do this is to counter those technological front-running strategies,” he says. Those with the means and the speed have such an advantage over other institutional investors, not to mention retail investors, he maintains, that disarming bad actors in some form is necessary to restore balance and fairness.
But just how big of a problem is front-running today? No one can say for certain. “Front-running by HFT has never been definitively proven,” says Andriy Shkilko, an associate professor of finance at Wilfrid Laurier University who researches HFT. Studies that have tried to quantify it were inconclusive. “There’s actually no reason to believe that technological front-running is a big thing.” Institutional investors have had years to adapt to the presence of HFTs, and there are many smart-order routers on the market that decide how to break up large orders of stock across various markets so that HFTs have a harder time anticipating trades. If investors are smart about how they execute trades, Shkilko says, “front-running is very difficult.”
There are other harmful strategies attributed to HFT, however, such as spoofing. That’s where a trader places a bunch of buy and sell orders with no intention of completing them. Spoofing can trick another market participant into buying shares at a higher price, for example. The Investment Industry Regulatory Organization of Canada considers spoofing to be manipulative and deceptive, but it’s hard to prove malicious intent, since there are legitimate reasons to cancel an order. Still, Shkilko says there are problems like spoofing that Aequitas’s platform can potentially address, even if there’s no “hard-core evidence” showing these practices are widespread.
RBC’s Mills recognizes how difficult it is to quantify HFT’s impact, but he doesn’t see that as a reason for inaction. “The academics who try to understand this, I give them great marks. But the fact of the matter is they’re never going to have complete information,” he says, adding, “none of them sit on trading desks.” Those who do, Mills says, know their trading costs have gone up since HFT became a force in the markets and also have to contend with the costs of constantly upgrading their technology to try to stay competitive. The uncertainty around the impact of HFT might weaken Aequitas’s reason for being, but CI Financial’s Sunseri says the Neo platform provides certainty. “It’s a safe harbour,” he says. “The first place we’re going to hit is Aequitas, and we’ll probably leave standing orders there, because we’re pretty confident we’re not going to be sniffed out by the high-frequency guys.”
CI is a backer of Aequitas, of course. But will other market players use it? Aequitas does have a formidable competitor in TMX Group, which will fight any loss of market share with all the tools at its disposal. In fact, TMX has already introduced features designed to address some of the perceived issues around HFT. “We hear from our customers that certain things are challenging, and that’s led us to our equities repositioning,” says Cowan. On the Alpha exchange, the company reversed a fee structure known as the maker-taker model. Previously, firms were paid a rebate by the exchange for providing liquidity and charged fees for taking advantage of it. The model has come under fire for fuelling another HFT strategy known as rebate arbitrage, which some believe distorts pricing and hurts liquidity. The company is also introducing a speed bump to Alpha. (A proposal is currently before the OSC.) The delay applies to all investors, unlike that of Aequitas. “We think the other approach is fundamentally unfair, and that it’s not a good example of transparent and open access,” Cowan says. “There are lots of examples in the commercial world, the political world and everywhere else where somebody has the arbitrary discretion to draw distinctions between different groups of people, and that inevitably leads to unfairness and lack of transparency.”
Schmitt, in turn, argues the TMX’s speed bump is misleading. There is one order type to which it does not apply, known as post-only: It allows traders to enter an order without having to complete it. “They have an opportunity to look at what’s really happening in the market, and if they like it, they stay. If they don’t, they move away,” Schmitt says of Alpha’s approach. “This is a model that is purely built to support high-frequency trading.” (The TMX disagrees for a few reasons, one of which is that it expects orders to be fully filled on Alpha without the need to send them to other markets, so HFTs would be unlikely to withdraw liquidity elsewhere as a result.)
But the time for trash-talking is drawing to an end. Soon Schmitt will have to make good on his word. He isn’t that concerned with capturing a lot of market share out of the gate, he says, but has loftier ambitions to reduce the cost of capital, foster new companies and ultimately increase the equities pool in Canada as a whole. Schmitt points to the recent changes at the TMX as proof that Aequitas is already having an impact. “That means we’re doing something right,” he says. He can’t resist one last dig.
MORE ABOUT INVESTING: