Faceoff: High frequency trading

By Jessica Bruno | December 10, 2013 | Last updated on December 10, 2013
9 min read

Doug Clark, Managing Director, Research, ITG Canada Corp

Stance: Overall, HFT doesn’t matter

HF trading barely hits retail investors

If you’re going to buy 300 shares of TD, it has near-zero impact on you. Where it does impact the retail investor is in mutual funds or pension plans that are trying to buy more stocks. That’s where high-frequency (HF) [traders] are making most of their money. They’re catching signals, figuring out that there’s a large buyer of XYZ or a large seller, getting in front and making those trades more expensive.

The exchanges have made it advantageous for those who can react very quickly. For example, when I try to buy 10,000 shares [across] three markets, if I don’t do an extremely good job of it, I won’t get all 10,000. If I get to one market too quickly, the others will disappear because the high-frequency guys will either pull their own offers or lift that stock in front of me.

Retail investors’ orders are so small that they’re typically trading them on one market. The investors would be affected to a small extent, but for the most part, it’s only when they’re trading via a mutual fund or the like that the orders are big enough.

Technology costs

There are products I would love to develop more quickly for my clients that I can’t because the TMX is forcing me to update my trading systems. It’s a constant technology demand from the marketplace, and at no point has a regulator said, “This is too much for the dealers.” For the mid-tier, or the less technically capable dealers, it’s more than they have resources for. So it really limits your ability to build risk management tools, or to improve your algorithms, for example. Last year, we spent millions updating our tech. Most of the traditional firms in Canada are not spending the money on technology that we are. As a result, some 20% or 30% of the time they trade orders, they’re not getting what they expect to get. This affects firms that are trading for the institutions.

The 2010 flash crash

It will happen again. There have been flash crashes before we had high-frequency. There was a similar market event in 1962, long before we had electronic trading, where markets lost 20% in 20 minutes and then bounced back.

If you owned a stock on the day of the 2010 flash crash and you went to work and didn’t look at a quote, it was trading roughly equal to where it was the day before by the end of the day. Did it really impact you? Probably not—the real impact is on market confidence. Do people walk away from trading in this market because they no longer understand it? Yes they do. The new market Aequitas, for example, surveyed the investment advisor community. They were significantly underwhelming in terms of their confidence in equity markets.

Hidden costs

In Canada, [about] a dozen dealers resigned their memberships this year, just because they can’t afford to continue to operate in the marketplace. Union Securities and Stifel Nicolaus walked away from Canada. There’s going to be a cost to the end users, whether it’s less choice among dealers, less research, or less technology available to them. It’s tough to increase commissions, because the first person to try and do so is going to lose business.

There are a lot of academic studies that say, “HF trading is great because the market cost has gone down.” But they don’t consider—simply because they don’t have the data—what the costs are to the dealer community, what the opportunity costs are, what the various other frictions are, and the hidden costs of maker-taker pricing (see “What is maker-taker pricing?”, this page).

What is maker-taker pricing?

Over the past 15 years, exchanges have changed their trading fees. Under the old system, exchanges charged a small fee to the buyer, the seller or both. Now exchanges charge a single, higher fee to the person who initiates a trade by putting in an order (known as the taker), and rebate part of that fee to the liquidity supplier, who is considered a market maker. A typical fee for the taker is 30 cents, while the maker gets a typical rebate of 25 cents. The exchange pockets the difference, and it’s estimated exchanges are making the same profit per trade as they were under the old system. In U.S. equity markets, the amount of fees going from takers to makers is about $2 billion a year.

—Source: “Maker-Taker Pricing Effects on Market Quotations,” Larry Harris, USC Marshall School of Business.

Exchange favouritism

Let’s say you place your [1,000-share] order on one market place, and I place my order on another marketplace. A high-frequency trader is going to link our two orders, and now 2,000 shares have been traded. The TMX gets paid on 2,000 instead of 1,000. The trader has exacted a little bit of a profit on it. But you and I aren’t doing any better. The dealer who is handling our orders is paying a higher fee to do so. It’s a tough argument to say the high-frequency traders are needed. They’ve basically created an opportunity for the TMX to run three markets.

When you create multiple exchanges under one umbrella group, they don’t really compete with each other. They’re not going to steal market share from one another; they work in a way that creates volume between them. That’s not competition, that’s fragmentation. The marketplaces have to get to a level playing field where the exchanges don’t feel they should be treating one set of participants better. When everybody paid the same fee—you paid 0.02 cents per share when you traded, whether you were active or passive—everybody was playing for the same economics.

Traditionally, the marketplace only had a buy order, a sell order and a short-sell order. In the U.S., some exchanges have dozens of orders, and permit all ways of using orders. So, you actually get up to thousands of possible ways to use an order. All these order types help the guys who are doing rebate arbitrage [on passive trading], and you never create an order to help the other side of the trade.

As soon as you start favouring these passive behaviours, you incentivize behaviours where there are high-frequency trading firms that buy and sell stock at the exact same price all day long just for these rebates. Their trades are meant just to get in the way. They only trade the most active names, and the liquidity they’re adding is of very limited value.

It may not even change the price: the price on Bombardier may be $3.50 to $3.51 all day long but as a natural retail or institutional buyer, you no longer can buy it on the bid; you have to cross the spread. That alone can have a market impact, even if the quote of the stock doesn’t move at all.

Should advisors worry?

You’re not going to see regulatory solutions. You’re going to see commercial solutions. You are starting to see a bit of pushback [with] markets like Aequitas in Canada (there’s a similar market in the U.S. called IEX). What’s going to happen is they’re either going to be successful and force the NYSE to change its behaviour, or they’re not, which will tell you nobody is that disadvantaged that they’re going to change.

Spend more time worrying about what you’re actually investing in rather than the nuance of how this is going on. But where you should be concerned is if this results in people putting money into mattresses rather than putting money into markets. Not only is that not good for the advisor, it’s not good for the economy.

Kevin Sampson, Vice president, business and strategy, TMX

Stance: HF trading helps investors

HF trading reduces costs

We see benefits to the market in the form of reduced spreads. Ultimately, investors are getting better prices. Our markets are probably as efficient, or more efficient, now than they have ever been. We hear from customers that [higher costs due to HF trading activity] do happen on occasion, but that’s a challenge that will always exist. It’s incumbent upon us to try to understand what those challenges are.

Policing behaviour

The focus should be on what behaviour we think is contrary to the interests of having a fair, orderly market—whoever is conducting that type of behaviour, whether it’s an HF trader or another group. If you’re putting up a quote with no intention to buy or sell the security, that’s contrary to trading rules in Canada (see the IIROC’s Guidance on Certain Manipulative and Deceptive Trading Practices).

The 2010 flash crash

They’re isolated and rare events. It’s been shown through studies that have been done in the U.S. that it wasn’t predominantly HF traders that caused the flash crash, for example, or the Facebook IPO. But it goes back to the perception. The investor just gets this feeling that something doesn’t look right or isn’t functioning properly. That influences the degree to which they participate in the markets. That’s why it’s incumbent upon us, and all market participants, to educate people. We need to make sure we’re talking facts versus myths or fear-mongering.

Exchange favouritism

We certainly reject the claim [that we favour HF traders]. We’re always going to look at where there’s opportunity to better serve our diverse customer base, and just because we develop or introduce a product or service that might be of disproportional value to one constituency over another, doesn’t mean we’re not servicing the other constituency.

We service a lot of different types of investors and trading participants. … We see 15% to 20% [of transactions] tops that may be attributed to HF trading, depending on the day. That’s a small percentage, certainly in comparison to other jurisdictions. Editor’s note: according to the C.D. Howe Institute, HF trading accounts for 50% of trades in the U.S.

Technology costs

Back in the floor trading days, whoever was closest to the post had an advantage. People will always be looking for and deploying more advanced technology in order to facilitate trading. It was through one method decades ago, and today’s new method is high-performance market-data feed readers, sophisticated fixed gateways and proximity services, and that’s going to continue to evolve.

People are going to always have a choice on what they want to spend and how important it is to them to have efficient and effective technology. We want to ensure that we’re not the slow link in the chain. As more technology becomes commoditized and the price points come down, that’s to the benefit of everybody.

If you look at our average fee per share from a transaction perspective, in the last 15 years, it’s a pretty dramatic decrease. We have extended those cost savings over time to our customer base.

But it also puts us in a good position to adopt some of that technology and be able to extend it to our broader customer base, where still some of those customers may not be able to independently absorb those costs themselves.

Should advisors worry?

For a lot of people, their investment strategies and how they service their clients have very little to do with having high-technology capabilities.

If you’re a long-only, buy-and-hold type of investor, you just want to see the market appreciate over time. You don’t need high-technology capabilities to meet your investment objectives.

A proprietary trader’s take

Ashraf Yaghi, Director, Vortex Capital Group

Does the money needed to build a HF trading infrastructure unfairly exclude some players?

Today, the cost of HF trading infrastructure has dropped mainly due to increased competition from different providers. Many players that didn't have access are now able to obtain it at a fraction of the cost from just a few years ago.

A large part of the success of HFT depends on speed, which in turn depends on co-location. A prime example of co-location service discounts come from Nasdaq OMX, which is trying to lure customers by offering cost savings as high as 50%.

Do you think HF trading will eventually become the trading standard, or will it continue to exist alongside traditional trading methods?

At its peak a few years ago, it almost accounted for 70% of trading volumes.

That’s since fallen to approximately 50%. A major trend has been the shift to emerging market exchanges, such as the BM&F Bovespa (Brazil) and Bolsa Mexicana (Mexico), which are upgrading execution infrastructures to support HF trading.

Jessica Bruno