Flash Boys: A Wall Street Revolt

Chapter 1 begins in the summer of 2009 with a crew of 2000 workers laying an underground cable from Chicago to New Jersey in the straightest line possible. This massive effort was undertaken in secrecy. The workers knew nothing about why they were blasting tunnels and laying down the line. The project was the brainchild of Dan Spivey, a Mississippian with a background in the Chicago stock exchanges. Spivey believed that Chicago exchanges could gain a competitive advantage by shortening the time it took to complete a trade with the Wall Street exchanges in New York. Even a difference of a few milliseconds, one-thousandth of a second, could allow one exchange to execute a trade ahead of their rivals. Computer-based trading made it possible to compete within timeframes that the human mind could barely perceive. Existing communication lines between brokerages were inefficient because they ran along railroads or bounced from one population center to the next. Spivey’s alternate route between Chicago and New Jersey was more than 100 miles shorter and avoided pedestrian digital traffic. Spivey consulted a group of engineers and local government officials to figure out the physical and legal logistics of laying down such a line, and secured financial backing from Jim Barksdale, the former CEO of the internet company Netscape Communications. Barksdale and Spivey formed Spread Networks and began construction of the line in early 2009. They encountered many difficulties along the way, from crossing through mountains and rivers to worrying about exposing their plans to better-financed rivals. Spivey assumed the line would be attractive to major Wall Street players—although this was based more on intuition than empirical evidence. Spread Networks decided they could sell access to 200 firms for a monthly rental fee of $300,000 as well as an up-front charge of about $14 million for a total of $2.8 billion to help subsidize construction.

By 2010, Spivey began making quiet pitches to Wall Street, insisting on strict secrecy with each potential client. It was hard to prove that the concept would work while the line was under construction, and Spivey could not explain what the benefits would be under the best of circumstances. Even so, the simple possibilities of losing out to a rival or increasing profits helped win over many prominent clients, though some resented how they could only use the line for their own transactions, not those of their clients. Just as they were about to secure their clientele, the line hit a major obstacle in the Susquehanna River, and local business owners in the nearby town of Sunbury, Pennsylvania were suspicious of outsiders promising big money for a project they wouldn’t describe. Spread Networks managed to overcome both sets of difficulties and completed the line at 827 miles, 13 miles short of their original projection.

Chapter 2 introduces Brad Katsuyama, a trader at the Royal Bank of Canada (RBC), a minor player compared to the major Wall Street banks in the early 2000s. Katsuyama moved to RBC’s New York offices in 2002, and the cultural differences shocked him. The frugality and everyday kindness that Katsuyama considered normal from his Toronto upbringing had little purchase in the ruthless, risk-accepting world of Wall Street. An exceptional student and athlete, Katsuyama became a stock trader because it suited both his analytical abilities and the camaraderie of team sports. In traditional Canadian fashion, RBC culture prized courtesy and integrity, pairing Katsuyama’s professional skills with his genteel personality. Problems began in 2006 when RBC acquired a US firm specializing in electronic stock trading. Its new American employees brought in their own culture of chauvinism and willingness to ignore the law. RBC bought the firm assuming that electronic trading was the future, and while Katsuyama was skeptical that his new co-workers had the skills to fulfill this potential, he understood the significance of electronic trading. In 2007, Katsuyama noticed that when he clicked a button to buy a batch of stocks, the offers would either vanish or introduce new terms rather than process the sale at the advertised price. This meant that Katsuyama, whose job was to navigate markets for his clients, could no longer trust the information on his screen as an accurate representation of the digital-only market. He once sold a large volume of stock only to find the price of the stock plummeted as he finalized the sale, resulting in immense losses.

At first, everyone around Katsuyama assumed he was making a mistake. When it was apparent that markets were reacting to his actions, the software developers were stumped. Part of the problem likely stemmed from the rapid expansion in the number of digital-only exchanges since 2005. Complex algorithms could route trades in any number of different directions with variations in speed, since the major telecommunications carriers had not prioritized differences in milliseconds. These various exchanges also offered contradictory incentive structures, with some rewarding the trader for buying a stock beneath its advertised price and another punishing it for doing the same thing. This created incongruent results between different exchanges even when the trades themselves were similar. Katsuyama came to the distressing realization that new players were popping up and accruing enormous influence in this digital-only market, while knowledge of these new players often escaped the traditional powers on Wall Street. Confused and exhausted, Katsuyama nearly left the business until RBC asked him to take charge of electronic trading. He committed to figuring out the problem and realized that it went far beyond RBC. He lacked the technical knowledge to find the solution himself, so he recruited fellow Canadian Rob Park and a small group of computer programmers to help him. With RBC’s permission to lose $10,000 a day, they began testing a set of theories. The first theory was that the exchanges had developed a sequencing algorithm to prioritize certain kinds of orders over others, so that less desirable transactions became “phantom orders” when someone tried to execute them. This theory fell apart whenever Katsuyama’s team carried out an order with a single exchange. Katsuyama then realized that the most reliable exchange was also the one physically closest to his office, leading him to wonder if the incredibly small differences in time required to send signals to more distant exchanges could make a difference.

Katsuyama’s team wrote a program that slowed the delivery times to closer exchanges, so that their orders all arrived at the same time, preventing the more distant ones from reacting to movement in the closer exchanges. The program worked, and they called it THOR (Tactical Hybrid Order Router). THOR revealed that the “ticker tape” image of the stock market had become obsolete despite its relevance in the popular imagination. Computers could distort markets within thousandths of a second so that barely anyone noticed. The chief beneficiaries of this system were called high-frequency traders. Katsuyama wanted to blow the whistle on this market manipulation, but he needed somebody to report on HFTs and their activities from the inside.