High-Speed Trading Isn't About Efficiency—It's About Cheating

When hedge funds use bots to buy and sell stocks within milliseconds, they're not improving the market. They're rigging the market.

Greed is good, and high-speed trading is better—if you want to rig the market (Wikimedia Commons)

I wasn't sure I'd heard him right.

I was a senior in high school, and I was staring at NBA legend Red Auerbach. He'd coached the Boston Celtics to nine championships in 10 years, won seven more as an executive, and, a bit less notably, gotten his first coaching job at our school way back when. He was 85 years old, but he lived nearby and had finally agreed to come back to be feted.

We piled into the gym and buzzed as Auerbach ascended the make-shift stage at center court. There were introductions and congratulations and then it was his turn to talk. He was old, but still sharp. He regaled us with an embellished, if not apocryphal, story about how his proudest coaching victory had come at our school. That was back in 1941, and the score had been something like 10 to 8. There was also something about yelling at the son of a senator—this was a preppy, all-boys school in Washington D.C.—for trying an around-the-back pass.

Then Auerbach turned to life lessons. "Everybody always asks me how to gain a competitive edge," he said, "and I'm always surprised because the answer is so obvious." Eighteen-year old me knew where this was going. He was going to tell us to work hard, that successful people prepare for their luck, yada, yada, yada.

"You cheat."

Our teachers looked confused, then horrified. They kept waiting for Auerbach to say he was just kidding, that of course there's no substitute for hard work. He didn't. Instead, he calmly explained that if you're playing a better fast-breaking team, you should install nets so tight that the ball gets stuck. Or if you're playing a faster baseball team, you should water the basepaths till they turn into muddy quagmires that nobody can run on. But most of all, he wanted to make sure we didn't misunderstand him. He cleared his throat, and said, "So, if you want a competitive edge, just cheat." Then he walked off stage, and the mayor's mother, who was inexplicably there, led us in a solemn rendition of America the Beautiful.

That brings us to high-frequency trading (HFT) hedge funds. These funds use computer algorithms—a.k.a.: algobots—to buy and sell stocks at incredible speeds. We're talking milliseconds. The idea is to react to any market news or inefficiencies before actual humans can process them. And it's an idea that has taken over stock trading: algobots make up about half of all stock transactions in 2012 (which is actually down from its peak of 61 percent).

It's Wall Street at its most socially useless. HFT funds aren't allocating capital to where they think it'll be most productive. HFT funds are allocating capital to where they think other people will put it 50 milliseconds from now. It's a tax on everybody else. And it's a tax that has basically no benefit. Sure, HFT funds defend themselves by saying they're increasing liquidity, but increasing liquidity is the last refuge of bullshitters. Just look at the chart to the left from Felix Salmon. It shows that the cost of trading has fallen as our computerized markets have become more liquid, but almost all of the drop happened before HFT. Economist Paul Samuelson had it right all the way back in 1957: knowing (or trading) something one second before everyone else is personally profitable and socially pointless.

And it's becoming more pointless now that markets are an algobot battleground. HFT funds aren't trading as much anymore, because there aren't enough humans to trade with. Algobots just quote each other prices. As Salmon points out, there were 280,000 quotes in a 17-minute span to trade the EFZ back in September 2012. But there were zero actual trades during that time. It's no surprise then that HFT funds are desperate for any kind of competitive advantage. So, like Red Auerbach suggested, they cheat. Now, what they do is strictly legal, but that doesn't make the game any less rigged. The Wall Street Journal reports that HFT funds buy early access to data from third-party distributors—everything from corporate earnings to the Philadelphia Fed's manufacturing survey. They're getting the numbers just fractions of a second early, but that's more than enough in the world of high-frequency trading.

There's a big difference between buying early access to public data and early access to private data. The University of Michigan, for example, sells the rights to its Survey of Consumers to Reuters for $1 million a year. Reuters then sells early access to it either five minutes before the public gets it or five minutes and two second before—for the HFT crowd that wants to frontrun the frontrunners. It's a horribly unlevel playing field (and we should tax some of it away), but, as Matt Levine points out, the University of Michigan might stop doing the survey if they couldn't make money off it.

But earnings reports and Fed data are different. The private sector isn't paying for the creation of a public good when it buys a sneak peek at them. The private sector is just profiting off existing public goods. If a company sold hedge funds an early look at their earnings, it'd be insider trading. But when a third-party like Business Wire sells hedge funds an early, albeit split-second, look at corporate earnings, it's perfectly legal. It's nuts. As Paul Krugman argues, we need a financial transactions tax—something like 0.1 percent on all trades—to make this kind of socially useless speculation personally useless too. Long-term investors wouldn't notice this small a tax, but ultra-short-term investors would: their warp speed trading would become less profitable and less prevalent.

I'm pretty sure—or at least I hope—that Red Auerbach was kidding when he told a gym full of kids to cheat to get ahead. But it's how too much of Wall Street operates today. And it makes a mockery of the idea that they've "earned" their wealth.