Putting a Speed Limit on the Stock Market

When Brad Katsuyama was running the U.S. trading desk for the Royal Bank of Canada, his clients would send in orders every day, but every day, when Katsuyama went to buy or sell, something would go wrong. When he wanted to buy, offers to sell shares would suddenly vanish, and the price of the stock would shoot up. When he wanted to sell, the same thing would happen in reverse. “I started to realize that, day in and day out, I was getting screwed,” Katsuyama told me recently.

The problem was that he was often too slow. Back then, in 2007, the stock market was in the middle of a significant shift. A combination of new technology and new regulations had led to the rise of firms focused on high-speed, computer-driven operations known as high-frequency trading. With the help of complex algorithms and ultrafast Internet connections, the new traders could buy and sell stocks in fractions of seconds, looking to make a seemingly infinite number of quick, tiny profits that added up. By 2009, high-frequency traders were making billions of dollars a year, and their transactions accounted for about 60 percent of U.S. stock trades.

Some of these traders acted like useful stock-market middlemen, constantly buying and selling, bridging the bid-ask gap between other buyers and sellers. But plenty of others used the new technology to foil long-term investors by trading ahead of the slower players. A trader’s algorithm might detect that Katsuyama was trying to buy 100,000 shares of a stock and then immediately start buying it to drive up the price. Indeed, certain high-frequency traders were forcing long-term investors, including those who managed funds that held ordinary people’s retirement accounts, to constantly buy higher and sell lower. The game seemed rigged.

At first, Katsuyama responded by creating an algorithm intended to make it harder for high-frequency traders to race in front of his trades. But then, he told me, he realized his clients’ real problem was not the traders themselves; it was the stock exchanges. As high-frequency traders proliferated, these platforms were adding clever services to attract their business. “If you want to solve the problem, you go to its root,” Katsuyama said. “And at the root, the problem is the market.” So last year, he and a few of his colleagues decided to leave the bank and start a new place for investors to trade. Rather than woo high-frequency traders, they would limit their advantages. Their trading platform, IEX, is set to open later this month.

The rise of high-frequency trading is often told as a technology story. Hedge funds, Wall Street banks and other firms used increasing computing power to write ever-smarter, ever-faster trading algorithms; fantastically expensive fiber-optic lines were built to increase transaction times by milliseconds. But the rise of high-frequency trading is also a result of the unintended consequences of regulation. Back in the ’90s and mid-aughts, a series of S.E.C. rules were designed to help ordinary investors by forcing stock exchanges to compete against one another. Exchanges could no longer hoard orders; if a better offer existed on another trading platform, they’d have to send it there.

It was a well-intentioned idea designed to better match buyers and sellers, but it wound up complicating things. New exchanges quickly arose to attract customers, and what is blandly referred to as the stock market soon became a complex web of more than a dozen separate exchanges. Today, it’s not just the familiar New York Stock Exchange and Nasdaq, but also lesser-known ones with names like BATS and Direct Edge. Exchanges are now basically just technology companies, rooms full of computers that match buyers and sellers. There are also roughly 50 so-called dark pools, which allow traders to trade the same stocks that change hands on ordinary exchanges but don’t require participants to post as much information. (IEX will be a dark pool with plans to grow into a full-fledged exchange.) “I don’t think anybody who was putting these rules together envisioned we would have 13 exchanges,” Charles Jones, a finance professor at Columbia Business School, told me.

Jacob Goldstein is a reporter for NPR’s “Planet Money,” a podcast and blog. Adam Davidson is off this week.

A version of this article appears in print on October 13, 2013, on page MM14 of the Sunday Magazine with the headline: TRADING PLACES. \n \n\n'; } s += '\n\n\n'; document.write(s); return; } google_ad_output = 'js'; google_max_num_ads = '3'; google_ad_client = 'nytimes_blogs'; google_safe = 'high'; google_targeting = 'site_content'; google_hints = nyt_google_hints; google_ad_channel = nyt_google_ad_channel; if (window.nyt_google_count) { google_skip = nyt_google_count; } // -->