Flash Boys: A Wall Street Revolt

by Michael Lewis

  Another investor, silent till that point, now raised his hand. “It seems like there’s a first mover risk for someone to behave the right way,” he said. He was right: Even the banks that were behaving relatively well weren’t behaving all that well. A big Wall Street bank that gave IEX an honest shot to execute its customers’ orders would suffer a collapse in its dark pool trading, and in its profits. The bad banks would pounce on the good bank and argue that, because its dark pool was worse than all the others, it shouldn’t be given the orders in the first place. That, Brad told the investors, had been maybe his biggest concern. Would any big Wall Street bank have the ability to see a few years down the road, and summon the nerve to go first? Then he clicked on a slide. On top it read: December 19, 2013.

  YOU COULD NEVER say for sure exactly what was going on inside one of the big Wall Street banks, but it was a mistake to think of a bank as a coherent entity. They were fractious, and intensely political. Most everyone might be thinking mainly about his year-end bonus, but that didn’t mean there wasn’t one person who wasn’t, and it certainly didn’t mean that everyone inside a big bank shared the same incentives. A dollar in one guy’s pocket was, in some places, a dollar out of another’s. For instance, the guys in the prop group who traded against the firm’s customers in the dark pool would naturally feel a different concern for those customers than the guy whose job it was to sell them stuff would—if for no other reason than that it is harder to rip off a person when you actually need to see him, face to face. That’s why the banks kept the prop traders on different floors from the salespeople, often in entirely different buildings. It wasn’t simply to please the regulators; all involved would prefer that there be no conversation between the two groups. The customer guy was better at his job—and had deniability—if he remained oblivious to whatever the prop guy was up to. The frantic stupidity of Wall Street’s stock order routers and algorithms was simply an extension into the computer of the willful ignorance of its salespeople.

  Brad’s job, as he saw it, was to force the argument between the salespeople and the prop people—and to arm the salespeople with a really great argument, which included the distinct possibility that investors in the stock market were about to wake up to what was being done to them, and go to war against the people who were doing it. In most cases, he had no idea if he had succeeded and, as a result, suspected he had not.

  Right from the start, the view from inside Goldman Sachs had been less cluttered than the view from inside the other big Wall Street banks. Goldman was unlike the other banks; for instance, the first thing the people he met at the other banks usually did was tell him of the hostility all the other banks felt toward IEX, and of the nefariousness of the other banks’ dark pools. Goldman was aloof, and didn’t appear to care what its competitors were saying or thinking about IEX. In their stock market trading and perhaps in other departments as well, Goldman was undergoing some kind of transition. In February 2013, its head of electronic trading, Greg Tusar, had left to work for Getco, the big high-frequency trading firm. The two partners then assigned to figure out Goldman’s role in the global stock markets—Ron Morgan and Brian Levine—were not high-frequency trading types. They didn’t bear a great deal of responsibility for whatever the high-frequency trading types had done before they took over. Morgan worked in New York and was in charge of sales; Levine, responsible for trading, worked in London. Both were apparently worried about what they had found when they stepped into their new positions. Brad knew this because, oddly, Ron Morgan had called him. “He found us by talking to clients about what they wanted,” said Brad. A week after they first met, Morgan invited Brad back to meet with a group of even more senior executives. “That didn’t happen anywhere else,” Brad said. After he left, he was told that the ensuing discussion had reached “the highest levels of the firm.”

  In taking over, Morgan and Levine had been tasked with answering a big question posed by the people who ran Goldman Sachs: Why was Morgan Stanley growing so fast? Their rival’s market share was booming, while Goldman’s was stagnant. Levine and Morgan did what everyone on Wall Street did when they wanted to find out what was going on inside a rival bank: They invited some of its employees in for job interviews. The Morgan Stanley employees explained to them that the firm was now trading 300 million shares a day—30 percent of the volume of the New York Stock Exchange—through what it called “Speedway.” Speedway was a service Morgan Stanley provided to high-frequency traders. Morgan Stanley built a high-frequency trading infrastructure—co-location at various exchanges, the fastest routes between them, a straight road into the bank’s dark pool and so on—and then turned around and leased their facilities to the smaller HFT firms, which couldn’t afford the up-front cost of building their own systems. Morgan Stanley got credit for, and commissions from, everything the HFT guys did inside Morgan Stanley’s pipes. The Morgan Stanley employees angling for jobs at Goldman Sachs told the Goldman executives that Speedway was now making Morgan Stanley $500 million a year, and that it was growing. This raised the obvious question for Goldman Sachs: Should we create our own Speedway? Should we further embrace high-frequency trading?

  One of Goldman’s clients handed Ronnie Morgan a list of thirty-three big investors to whom he should speak before making this decision. This client didn’t know if Morgan had spoken to people beyond this list, but he confirmed for himself that Morgan had spoken to each of the thirty-three people individually. At the same time, Morgan and Levine began to ask some obvious questions about Goldman Sachs’s stock market businesses. Could Goldman ever be as fast or as smart as the more nimble high-frequency trading firms? Why, if Goldman only controlled 8 percent of all stock market orders, was it able to trade more than a third of those orders in its own dark pool? Given how little of the flow Goldman saw, what was the likelihood that the best price for an investor’s order came from some other Goldman customer? How did Wall Street dark pools interact with each other and with the exchanges? How stable was this increasingly complex financial market? Was it a good thing that the U.S. stock market model had been exported to other countries and other financial markets?

  They already knew or could guess most of the answers; for the questions still hanging, the investors pointed them toward an unusually forthright and knowledgeable guy they knew and trusted who was starting a new stock exchange: Brad Katsuyama.

  What struck Brad about his visit to Goldman Sachs was not only that Levine and Morgan were willing to spend time with him, but that they took the ideas from their conversations to their superiors. Levine seemed particularly concerned about the stock market’s instability. “Unless there are some changes, there’s going to be a massive crash,” he said, “a flash crash times ten.” In conversation and in presentations, he impressed the point upon Goldman’s top executives, and also asked, “Do you really need the only differentiator in the market to be speed? Because that’s what it seems to be.” It wasn’t all that hard for the people who ran Goldman Sachs to see the source of the problem, or to see why no one inside the system cared to point it out. “There’s no upside in it—that’s why no one ever steps out on it,” said Levine. “And everyone’s got career risk. And no one is thinking that far ahead. They are looking at the next paycheck.”

  A long string of myopic decisions had created new risks in the U.S. stock market. Its complexity was just one manifestation of the problem, but in it, the Goldman partners both felt sure, lay some future calamity. The sensational technical glitches weren’t anomalies but symptoms. And a stock market calamity, Ron Morgan and Brian Levine both thought, would end up being blamed generally on the big Wall Street banks, and specifically on Goldman Sachs. Goldman earned $7 billion a year from its equity business; that business would be put at risk by any crisis.

  But it was more than that. At forty-eight and forty-three, respectively, Morgan and Levine were, by Wall Street standards, old guys. Morgan had been made a Goldman partner back in 2004, Levine in 2006. Both confided to friends
that IEX presented them with a choice, at what might be a pivotal financial historical moment. An investor who knew Ron Morgan said, “Ronnie’s saying to himself, ‘You work for twenty-five years in the business, how often do you have a chance to make a difference?’ ” Brian Levine himself said, “I think it’s a business decision. I also think it’s a moral decision. I think this is the shot we have. And I think Brad is the right guy. It’s the best odds we have to fix the problem.”

  BEFORE THEY OPENED their market, on October 25, 2013, the thirty-two employees of IEX made private guesses as to how many shares they’d trade their first day and in their first week. The median of the estimates came in at 159,500 shares the first day and 2.5 million shares the first week. The lowest estimate came from Matt Trudeau, the only one of them who had ever built a new stock market from scratch: 2,500 shares for the first day and 100,000 for the week. Of the ninety-four stock brokerage firms in various stages of agreeing to connect to IEX, most of them small outfits, only about fifteen were ready on the first day. “Brokers are telling their clients they’re connected, but we haven’t even gotten their paperwork,” said Brad. When asked how big the exchange might be at the end of the first year, Brad guessed, or perhaps hoped, that it would trade between 40 and 50 million shares a day.

  To cover their running costs, they needed to trade about 50 million shares a day. If they failed to cover their running costs, there was a question of how long they could last. “It’s binary,” said Don Bollerman. “Either we are a resounding success or we are a complete flop. We’re done in six to twelve months. In twelve months I know whether I need to look for a job.” Brad thought that their bid to create an example of a fair financial market—and maybe change Wall Street’s culture—could take longer and prove messier. He expected their first year to feel more like nineteenth-century trench warfare than a twenty-first-century drone strike. “We’re just collecting data,” he said. “You cannot make a case without data. And you don’t have data unless you have trades.” Even Brad agreed: “It’s over when we run out of money.”

  On the first day, they traded 568,524 shares. Most of the volume came from regional brokerage firms and Wall Street brokers that had no dark pools—the Royal Bank of Canada and Sanford Bernstein. Their first week, they traded a bit over 12 million shares. Each week after that, they grew slightly, until, in the third week of December, they were trading roughly 50 million shares each week. On Wednesday, December 18, they traded 11,827,232 shares. By then Goldman Sachs had connected to IEX, but its orders were arriving on the new exchange in the same untrusting spirit as those from the other big Wall Street banks: in tiny lot sizes, resting for just a few milliseconds, then leaving.

  The first different-looking stock market order sent by Goldman to IEX landed on December 19, 2013, at 3:09:42 p.m. 662 milliseconds, 361 microseconds, and 406 nanoseconds. Anyone who had been in IEX’s one-room office when it arrived would have known that something unusual was happening. The computer screens jitterbugged as the information flowed into the market in an entirely new way. One by one, the employees arose from their chairs; a few minutes into the surge, all but Zoran Perkov were on their feet. Then they began to shout.

  “We’re at fifteen million!” someone yelled, ten minutes into the surge. In the previous 331 minutes they had traded roughly 5 million shares.

  “Twenty million!”

  “Fucking Goldman Sachs!”

  “Thirty million!”

  The enthusiasm was unpracticed, almost unnatural. It was as if an oil well had gushed up through the floor during a meeting of the chess club.

  “We just passed AMEX,” shouted John Schwall, referring to the American Stock Exchange. “We’re ahead of AMEX in market share.”

  “And we gave them a one-hundred-and-twenty-year head start,” said Ronan, playing a little loose with history. Someone had given Ronan a $300 bottle of Champagne. He’d told Schwall that it had cost only forty bucks, because Schwall didn’t want anyone inside IEX accepting gifts of more than forty bucks from anyone outside of it. Now Ronan fished the contraband from under his desk and found some paper cups.

  Someone else put down a phone and said, “That was J.P. Morgan, asking, ‘What just happened?’ They say they may have to do something.”

  Don put down his phone. “That was Goldman. They say they aren’t even big. They’re coming big tomorrow.”

  “Forty million!”

  At his desk Zoran sat calmly, watching traffic patterns. “Don’t tell anyone, but we’re still bored,” he said. “This is nothing.”

  Fifty-one minutes after Goldman Sachs had given them their first honest shot at Wall Street customers’ stock market orders, the U.S. stock market closed. Brad walked off the floor and into a small office, enclosed by glass. He thought through what had just happened. “We needed one person to buy in and say, ‘You’re right,’ ” he said. “It means that Goldman Sachs agrees with us.” Then he thought some more. Goldman Sachs wasn’t a single entity; it was a bunch of people who didn’t always agree with each other. Two of these people had been given a new authority, and they had used it to take a different, longer-term approach than anyone imagined Goldman Sachs was capable of. These two people made all the difference. “I got lucky Brian is Brian and Ronnie is Ronnie,” said Brad. “This is because of them. Now the others can’t ignore this. They can’t marginalize it.” Then he blinked. “I could fucking cry now,” he said.

  He’d just been given a glimpse of the future—he felt certain of it. Goldman Sachs was insisting that the U.S. stock market needed to change, and that IEX was the place to change it. If Goldman Sachs was willing to acknowledge to investors that this new market was the best chance for fairness and stability, the other banks would be pressured to follow. The more orders that flowed onto IEX, the better the experience for investors, and the harder it would be for the banks to evade this new, fair market. At that moment, as Goldman’s orders flowed onto IEX, the stock market felt a bit like a river that wanted to jump its banks. All that had been needed was for one man with a shovel to dig a trench in an existing levee, and the pressure from the water would finish the job—which was why men caught digging into the banks on certain stretches of the Mississippi River were once shot on sight. Brad Katsuyama was the man with the shovel, positioned at the river’s most vulnerable bend. Goldman had arrived, with explosives, to help him.

  Three weeks later, he stood before a group of investors who, if they acted together, might force change upon Wall Street. To show them that change was possible, he flashed on a big screen the data from what had happened, for fifty-one minutes, on December 19. The data showed, among other things, the power of trust. Goldman had actually sent more orders to IEX the day before, on December 18. So much more had traded on December 19 because, on that day, for just fifty-one minutes, Goldman had entrusted them with most of its orders for ten seconds or more. That trust had been rewarded: The market felt fair; 92 percent of those orders traded at the midpoint—the fair price—compared to 17 percent that traded at the midpoint in Wall Street’s dark pools. (The number on the public exchanges was even lower.) Their average trade size was twice the market average, despite the efforts of other Wall Street banks to undermine them.

  IEX represented a choice. IEX also made a point: that this market which had become intentionally and overly complicated might be understood. That, to function properly, a free financial market didn’t need to be rigged in someone’s favor. It didn’t need in some sick way the kickbacks, and payment for order flow, and co-location, and all sorts of unfair advantages handed to a small handful of traders. All it needed was for the men in the room and other investors like them to take responsibility for understanding it, and then to seize its controls. “The backbone of the market is investors coming together to trade,” said Brad.

  When he was finished, an investor raised his hand. “They did it on December nineteenth,” he asked. “And then what?”

  _____________

  * Eric Hunsader, the
founder of Nanex, a stock market data company, is a fantastic exception to the general silence on this subject. After the flash crash, it occurred to him to use his data to investigate what had gone wrong, and the search never really ended. “Almost every rock I overturn, something nefarious crawls out from under it,” he said. Hunsader has brilliantly and relentlessly described market dysfunction and pointed out many strange micro-movements in stock prices. When the last history of high-frequency trading is written, Hunsader, like Joe Saluzzi and Sal Arnuk of Themis Trading, deserves a prominent place in it.

  † “Glitch” belongs in the same category as “liquidity” or, for that matter, “high-frequency trading.” All terms used to obscure rather than to clarify, and to put minds to early rest.

  ‡ From a book of that name by Charles Perrow.

  § In March 2013, the Commodity Futures Trading Commission, a derivatives regulator, ended its nascent program to give outside researchers access to market data after one of those researchers, Adam Clark-Joseph, of Harvard University, used the data to study the tactics of high-frequency traders. The commission shut down the research after lawyers for the Chicago Mercantile Exchange wrote the regulators a letter arguing that the data Clark-Joseph had collected belonged to the high-frequency traders, and that sharing it was illegal. Before he was booted out of the place, Clark-Joseph showed how HFT firms were able to predict price moves by using small loss-making stock market orders to glean information from other investors. They then used that information to place much bigger orders, the gains from which more than compensated for the losses.

  ¶ Estimates of commission paid to Wall Street banks for stock market trades in 2013 range from $9.3 billion (Greenwich Associates) to $13 billion (the Tabb Group).