High-frequency trading—often referred to as HFT—is a blanket term for computer-driven trading programs that now account for 60% or more of all stock market volume. Should we be afraid? We asked an expert.

“First of all, there’s no such thing as HFT,” says Bernard Donefer. “There are a number of different business models that use low-latency infrastructure, but you can’t just lump them all together as high-frequency trading and label them good or bad. You need to look at each business model individually and then ask, ‘does this benefit or hurt the market?’”

What makes Donefer an authority? First of all, he spent 35 years on Wall Street, most recently as the head of Capital Markets Systems at Fidelity Investments. Since retiring, he’s become Distinguished Lecturer and Associate Director of the Subotnick Financial Services Center at the Zicklin School at Baruch College, and Adjunct Associate Professor at the NYU Stern Graduate Business School.

Donefer has spent a long career studying the issue of automated trading, and as he points out, he no longer has a vested interest. “I’m not paid by anybody on either side. My job is just to explain things to the public,” he says. To hear the professor explain it, there are four main types of high-frequency trading that deserve scrutiny.

1. Algorithmic traders

“This is when traders on the buy side take a large order and cut it into many pieces to minimize the impact on the market. They’re not trying to hurt anybody, just lower their costs,” says Donefer. Indeed, breaking large trades into smaller pieces is something traders have done for decades. Except instead of having to hand-write dozens (or hundreds) of individual order tickets, traders now have algorithmic trading programs that can do the job automatically.

2. Automated market makers

A market maker’s job is to hold securities in inventory and display bid and ask prices to the market so that investors can find a buyer or seller for whatever they’re hoping to buy or sell. The market makers profit by pocketing the spread between the bid price and the ask price.

“Some fear that quantitative trading is unfair and could spiral out of control”

“Fifteen years ago, over-the-counter market makers were human beings, and they kept the spreads wide,” says Donefer. “One study showed they were making 25 cents per share on average for every trade. In 1997, the NASDAQ paid a $1 billion settlement for purposely keeping the spreads too wide.”

By contrast, today’s automated market makers have brought average spreads down to about a penny, and created liquidity for thousands of securities that were previously difficult to cost-effectively trade. Donefer sees automated market makers as a good thing—providing better liquidity and reducing conflicts of interest with investors.

3. Quants

Quantitative traders—or “quants”—use high-speed technology to exploit anomalies in the market. “One example is the pairs trade,” says Donefer. “Let’s say two pharmaceutical stocks normally trade closely together, then, all of a sudden, one jumps in price. When the computer sees this anomaly, it might short the stock that jumped and go long the other one, betting that the prices will re-converge. If this strategy works 60% of the time, the quant will make lots of money. If it works 40% of the time, he can become a college professor,” he chuckles.

Because these transactions can happen very quickly—perhaps hundreds of trades in a single second—some fear that quantitative trading is unfair and could spiral out of control.

Donefer disagrees: “Quants are using public data and are not violating a fiduciary duty to anyone else. They invest in low-latency architecture and need lots of brains to win. This is stuff we’ve always done, but it’s just faster with computers.”

4. Outlaws

The final group on Donefer’s list is the so-called outlaws—parties who are accused of using low-latency trading for dirty tricks such as “quote-stuffing” to bog down the market with bogus activity, and “spoofing,” where orders are entered with no expectation of being executed, just to convince others of widespread interest in a security.

“Show me someone using an improper technique to manipulate the market, and I say we should fine them and send them to jail,” says Donefer. “So far, I hear lots of talk and conspiracy theories, but I have yet to see any evidence. Just because there’s a gun on the table does not mean a murder has been committed.”

While Donefer concedes that high-frequency trading can cause problems in the market, he believes these problems are benign in nature, such as a bug in the software, a human error, or a delay or flaw in the market data. He also believes these problems will be identified and addressed with new protocols as they arise—much as the airline industry has succeeded in increasing safety year after year.

Ultimately, Donefer believes high-frequency trading is here to stay, and a good thing for investors of all stripes.

“You’re getting faster execution and smaller spreads than ever before. Costs of trading are lower than they’ve ever been. And if, like most investors, your goal is to hold a stock for days, weeks, months, or years, why do you care that a market maker made one-tenth of a penny on your trade?”

Our view

New technologies always introduce new challenges—whether it’s dealing with workers displaced by automated assembly lines, or figuring out how to protect privacy in the Internet age. While there are bound to be missteps along the way, we are hopeful that regulatory regimes will eventually catch up with high-frequency trading and guide its safe assimilation in capital markets.