It never was supposed to be this fast and furious.
When the first Wall Street traders hooked up their own personal computers (most likely IBM PCs or early IBM clones from Compaq), the machines were going to make things faster and easier -- more efficient, in other words. But soon, affordable networking cards attached the PCs to one another and then to the capital markets.
In the late 1990s and early 2000s, Wall Street went digital, and electronic trading exploded. Add in high-speed processors, low-latency networks, truly electronic trading platforms and dark pools, and the high-frequency future of trading had arrived.
But executing near the speed of light can be dangerous. Along with fast trades often have come loose control and severe crashes.
The majority of today's trades are executed electronically using algorithms; some estimates say algorithmic trading comprises as much as 74 percent of all stock trades in the United States -- and at faster speeds than once were imaginable. In the time it takes you to say "high-frequency trading," millions of shares in dozens of markets across the globe have changed hands. But a hyperefficient mechanism that is free of human interference, and therefore error, the current market is not.
Exceeding the Speed Limit
Regulators including the U.S. Securities and Exchange Commission and the Commodity Futures Trading Commission are struggling to keep up with and keep a grip on this warp-speed world -- and they are considering some severe measures as they struggle to understand just what high-frequency trading is. In a speech in February at New York Law School, CFTC commissioner Scott O'Malia admitted that his agency has failed to keep pace with the rapid evolution of trading technology.
This became apparent after the 2010 Flash Crash, when, O'Malia said, he realized the group lacked a fundamental understanding of the new trading patterns being used in the markets. "There was no consensus on what constituted automated trading, what constituted algorithmic trading and what constituted HFT," he told the audience. "There was no consensus on how each type of trading affected our markets. It was like having no consensus on what constitutes a broken window."
According to Mark Fickes, a former SEC senior trial counsel and currently an attorney with BraunHegey & Borden in San Francisco, the SEC is working to make sense of the new trading landscape. But it won't be cheap or pain-free, he says.
"One of the things that is most interesting, just based on the astronomical costs, is the proposal for the Consolidated Audit Trail, or CAT, that the SEC has been floating for a couple of years," says Fickes, referring to the real-time market data surveillance system. SEC chairwoman Mary Schapiro, who proposed the CAT initiative to Congress in May 2010 shortly after the Flash Crash, reported a staggering price tag for the system: about $4 billion to build and an additional $2 billion annually to maintain. "The SEC has been deliberately vague on the issue," Fickes notes.
The CAT proposal is further proof that the fallout of the 2010 Flash Crash and the collapse of MF Global is still reverberating throughout the industry. "Clearly what you are seeing is government regulators publicly acknowledging that there is a problem," Fickes says.
The HFT Debate
Although HFT has matured in the past decade, there are still fundamental arguments around what it is and what it does. At what speed does algorithmic trading become high-frequency trading? Did HFT cause the Flash Crash? Can traders possibly slow down now? And perhaps most important, is it safe?
Supporters of high-frequency trading wave the banner that they provide liquidity to a market that desperately needs it. Critics say HFT liquidity only drives volatility.
HFT may be a capitalistic pursuit that provides liquidity, but, "When you peel the onion, there is a flaw in the argument," says Michael Jenkins, a partner with global management consultancy A.T. Kearney. "High-speed trading leads to very high degrees of market fragmentation. Liquidity is only valuable when it can be aggregated and aggregated in a space where there's very high transparency and regulation that allows capital and ideas to trade freely." This is why we have high profile exchanges located in major cities, such as New York, London and Tokyo, he adds.
HFT is the opposite of that, Jenkins continues. "When we are down to the microsecond level -- one millionth of a second -- where we are now, traders hit a million markets every second, and this causes high degrees of fragmentation, which is the opposite of liquidity."
Garret Nenner of Momentum Trading Partners, a high-touch, agency-only broker-dealer that focuses on equities and options execution, also discounts HFT liquidity. "If you look at the HFT firms, you see a lot of prices are going in and they are testing the market. Those bids and offers are not real. They are real for the microsecond they are up there, but they are generally phantom orders," he says. "And that has to be recognized by the rest of the world."
Why? Because the vast majority of trades put forth by high-frequency traders are canceled immediately after they are placed. One expert estimates that 95 percent of all orders are canceled the instant after they are made.
"It's not trading; it's quoting," says A.T. Kearney's Jenkins. "It is market actions that are not about trading one security for a quantity of money. It's quoting and cancelations. It puts noise and volume into the market that do have some ability to distort what is going on."
At Advanced Trading's Buy-Side Trading Summit this year in Amelia Island, Fla., one panelist painted a vivid image of today's high-speed algo trader: "If a human trader behaved the way an algorithm does -- placing hundreds of orders and instantly canceling them -- there would be fistfights on the trading floor," Brooke Allen, head of the quantitative trading desk at Maple Securities, said. While the thought of an annoying trader with two black eyes and some missing teeth amused the audience, no one seems to disagree with the metaphor.