Market Mover

by Robert Greifeld

  But we still had work to do. As the economic headwinds of 2000–2003 slowly shifted to tailwinds, I knew it was more important than ever to prepare Nasdaq to thrive as the economy picked up speed. It was time to make a move that was not about getting slimmer, smarter, leaner, or even meaner, but was about positioning ourselves for future growth. In a narrow, cost-cutting context, it’s not easy to innovate and think long-term. When a culture is focused entirely on thrift, the next big thing is usually invented somewhere else. In other words, we couldn’t just play defense; we had to get our offense on the field as well.

  LEADERSHIP LESSONS

  • Prioritize Your Time to Get the Greatest Leverage. The CEO’s task list is endless, but time is not, so choose those activities that give maximum return on time spent.

  • You Can’t Do Everything Well. People sometimes think the key to success is doing your job well, but for a leader, it is equally important to know what you’re not going to do well and what you’re not going to do at all.

  • Run to Problems. Face reality relentlessly. Override the natural instinct to turn away from the things that are not working.

  • Develop Your Leadership Instinct. Find the right balance of experience, knowledge, data, and advice that works best for your decision-making process.

  • Don’t Underestimate Market Transitions. To win in the short term, you need to compete well against other players in your market, but to win in the long term, you have to get out ahead of major shifts in the market itself.

  Chapter Four

  Buy the Winners

  If you can’t beat ’em, buy ’em.

  “Nasdaq Agrees to Buy Brut ECN,” Wall Street Journal, May 26, 2004

  “Young man, you don’t get it. This is not an admission of failure.”

  The gentleman addressing me in this fatherly tone was Jim Mann, longtime CEO of SunGard. It was 1999, and he had just told me he was planning to acquire another company that had developed some technology we needed. I’d questioned his move—couldn’t we just build in-house the technology he was proposing to buy?

  “Give me a fraction of what you’re planning to pay for that company and I can come up with our own version,” I said.

  Mann shook his head, smiling indulgently at my entrepreneurial pride. “It’s smart,” he told me. “We buy the winners.” A successful company, he explained, has already beaten out its competition and proven itself in the marketplace while a dozen others failed. Yes, you pay a premium when you buy. But you save yourself all the time and resources it takes to develop your own product, build a customer base, and outperform your competitors. And you dramatically reduce your risk of failure by betting on a proven business.

  I took his words to heart. At forty-five, with many years in business under my belt, I was at an age where I could appreciate being called “young man.” And more important, I learned a lesson that day that I would employ for years to come: Buy the winners.

  Mann’s wise words came back to me as I assessed the state of Nasdaq’s technology in 2003 and 2004. Could we develop what we needed in-house, or should I be seriously looking for someone to buy?

  Turning Nasdaq into a market leader was going to require more than new tires and a paint job—we needed to overhaul the entire engine of our trading technology. I knew this before I was hired—that was why overhaul technology was one of the five points on the plan I presented to the Board. Even so, getting a look under the hood was an eye-opener. Nasdaq was operating on an older mainframe system, while the competition was using faster, more flexible, less expensive UNIX systems or, in some cases, Intel and Microsoft platforms. We had our own way of doing business that was native and proprietary, but the industry around us had become more open-source, fostering a degree of iteration foreign to our internal culture. If an ECN had a problem in executing orders, they’d just tweak it and reboot it, and their system would come back up—no harm, no foul. Like modern software, ECNs were constantly updating and upgrading, in some cases as often as once a day. Nasdaq had no such flexibility. If we went down, even momentarily, it would be on the front page of the Wall Street Journal. Our engineers upgraded the system closer to once a year. Yes, our platform was reliable—much more so than our new competitors’. The uptime in our trading systems in 2003 was more than 99.99 percent. In a different time and place, that reliability might have given us a significant advantage. But in this circumstance, it began to look like an albatross. It left us without the critical adaptability that allows innovation to thrive. As a result, we weren’t evolving fast enough. It was time for Nasdaq’s technology to enter a new era.

  I knew the technology behind electronic trading, and I had firsthand experience of building an ECN from the ground up. As I made my appraisal of the state of our technology in 2003, I took a good look at SuperMontage, knowing that the organization had invested enormous time and resources in developing this system. Could we turn it into a platform on which to build Nasdaq’s future? Try as I might, I just couldn’t see it.

  By 2004, we had begun to upgrade the existing transaction system to work with new regulations and keep competitive with existing ECN functionality. We were shifting to a more flexible, dynamic architecture and had even increased our average update speed from once a year to once a month or more if needed. Steve Randich, our CIO, was transforming our IT department into an outfit that valued dynamism and responsiveness, not just stability and reliability. BRUT had helped stabilize our market share. But I knew that none of those changes was sufficient to deal with our fundamental issues. With a technological engine designed to fit a different era, we couldn’t just put our foot on the accelerator.

  Realistically, that left two pathways open to us. Option one involved some promising but unproven new technology we had developed with Nasdaq Europe, built on a Microsoft platform. It could potentially be the architecture for Nasdaq’s next-generation system. But I was hesitant. It was not designed at enterprise scale, and I wasn’t convinced it would be easy to get it there. Speed to market was critical. The prospect of integrating a system, however promising, that wasn’t yet tested at scale and was bound to encounter growing pains along the way made me slightly queasy. We simply didn’t have a lot of margin for error. Spending too much time in development limbo or going to market with a less than bulletproof system was a fate to be avoided at all costs. Furthermore, I had done my previous system-building work on UNIX platforms, and knew the reliability of that computing ecosystem. I was willing to pursue the new technology, but our second option was my preferred path.

  What was option two? Simple: Buy someone. And not just anyone, but the best. Buy the winners. The winner, in my mind, of the ECN wars was very clear. It was INET (formerly Island ECN). Especially from a technological point of view, INET was the one I coveted. It was robust, sophisticated, and proven at scale. Its real genius was in the simplicity of its construction, often the hallmark of the best software architecture. It was elegant, efficient, and, with its off-the-shelf hardware and open-source software, had a light footprint that made it less expensive to operate than many of its competitors. Plus, it had the largest market share.

  To understand how INET technology had earned such a good reputation and why I singled it out, it helps to know something about its unlikely origin story—a remarkable tale starring a few highly motivated outsiders who were determined to disrupt the established Wall Street trading landscape with the new tools of the information revolution (and make a lot of money in the process). By chance, I’d made a brief appearance in one chapter of this narrative, more than a decade before I started the job at Nasdaq, in an unlikely setting—a basement office in Staten Island.

  An Island of Bandit Traders

  It was the early 1990s, long before the days of GPS navigation systems or ubiquitous mobile phones, and I was lost on Staten Island. I wasn’t familiar with the borough; like most New Yorkers, I’d only ever driven through. While Staten Island is just a short ferry ride from Wall Street’s soaring skyline, cul
turally it may as well be a thousand miles away. It is home to lots of working-class Irish and Italians, contains the highest number of gun owners in New York, and, in more recent years, boasted a majority of Trump voters. Some readers may remember it as the home of Melanie Griffith’s scrappy character in Working Girl. I thought about that movie as I anxiously cruised past blocks and blocks of identical row houses looking for the right address. As a blue-collar kid from Queens, I would have expected to feel at home. But the inimitable character of Staten Island resisted such superficial comparisons.

  The founder of ASC, Carl LaGrassa, had asked me to go in search of a trading outfit based somewhere out there. ASC had a back-office software business clearing trades. Most of the customers were small, doing a handful of trades every day. But suddenly, one customer’s trading volume spiked dramatically. “Find out what’s happening in Staten Island,” Carl had said.

  I was about to give up any hope of finding the place, when I turned a corner and saw a shiny Porsche, a Mercedes, and a BMW, all parked next to each other on the street. It was an incongruous sight, to say the least. At that time, in that neighborhood, it was like seeing a unicorn tethered on the side of the road. I had arrived.

  In the basement of that modest Staten Island home was the operation of Shelly Maschler. This was my first encounter with Maschler; it wouldn’t be my last. He was a fascinating character—a big, boisterous bear of a man, full of brains and braggadocio in equal parts. Think Rodney Dangerfield, only with a larger personality and more F bombs. Maschler would become something of a Wall Street legend in the years ahead, making millions by pioneering various trading schemes, many of which riled financial authorities, skirted the edges of regulations, and even ignored them altogether. In a few years, Maschler’s small Staten Island operation, Datek Securities, would morph into Datek Online, one of the original day-trading operations, which would eventually be sold to TD Ameritrade in 2002 for more than $1 billion.

  Maschler’s influence on Wall Street was significant, if controversial. Playing the role of the upstart, little-guy outsider fighting the elites, he used aggressive strategies to help usher in a new age of speed and automation in Wall Street trading practices. But perhaps his greatest impact was helping to recognize the talent and launch the career of Josh Levine, whom I met in his office that day.

  Levine was an idealistic, brilliant computer programmer who was just a few years out of high school and had landed a consulting gig working with Maschler and his small team of traders. Levine started out as a technical gofer—an enthusiastic kid still learning the finer details of trading and experimenting with technological enhancements that would help give Maschler’s team an edge. However, as the whole industry would discover in the years ahead, Levine had bigger goals. He believed in the promise of computers to change the world and, more important, to level the playing field on Wall Street.

  At the time we met, Levine was helping Datek capitalize on rule changes in the Small Order Execution System (SOES), an early computer ordering system that automatically processed the orders of small investors. SOES had been devised as a response to the infamous Black Monday crash of 1987, when smaller investors were furious that Nasdaq dealers didn’t answer their calls as stocks plummeted. From the dealers’ perspective, the unprecedented volume of trading that day meant that their phone lines were overwhelmed, and they simply couldn’t get to every call. Telephone, after all, is a nonscalable technology. Whatever the reason, the result spelled disaster for the small investor. So new rules were proposed to help protect them. The SOES system would automatically fill stock orders for small investors at a market maker’s current, posted quote (up to a thousand shares).

  As with many seemingly innocuous rule changes, the SOES rule spawned a new class of opportunities for savvy traders willing to exploit this automated system to their advantage (and ignore the spirit of the rule). Maschler was the most prominent of these “SOES bandits,” as the dealers called them. Given that any dealer or market maker worked in many Nasdaq stocks at once, and entered their posted stock quotes manually in their computer terminals, it was often hard to keep fully updated with small changes in each stock. At any given moment, two market makers might have different posted quotes, especially for stocks that were on the move. These were small differences, to be sure, but that was all a smart trader needed. For example, one dealer might have a posted quote of Microsoft at 25¾; another might have it at 26. Sitting in his basement in Staten Island, Maschler, or one of his traders, could buy from the first dealer and sell to the other in a split second, making money on the difference in price. One might make only tens or perhaps hundreds of dollars at most on any given trade, but repeat that many, many times a day, and pretty soon you are making real money.

  By any contemporary standard, it wasn’t technologically sophisticated. But when the other guy is unarmed, even the simplest weapon wins the day. And it wasn’t long until Levine devised a way to automate the process—a computer trading system that would carefully seek out the best trading opportunities on its own (it was appropriately called “the Watcher”). Other algorithms followed, and Levine’s reputation spread, helped along by another young trader in Maschler’s orbit, Jeff Citron. They showed me the system that day as I toured their Staten Island office. Watching their algorithm track “risers” and “fallers” among Nasdaq stocks, I suddenly wondered where they were getting this up-to-date information. There was no hard line attached to this house, as you might find on Manhattan trading desks. This was long before the days of mobile phones, high-bandwidth cable, or even dial-up internet. As I contemplated this seeming miracle, Maschler walked me outside and pointed up to the roof, where a satellite dish hung precariously.

  “Check out our data feed.” It was another anomaly in this sleepy neighborhood.

  “The neighbors must think you’re communicating with aliens,” I remarked, only half joking.

  With this innovative setup, Levine, Citron, and Maschler brought attention to the advantages that the smart use of new technologies could give the savvy (and sometimes unscrupulous) trader, increasing their trading speed and access to information. Soon, copycat strategies would be ubiquitous on Wall Street, even as they became increasingly sophisticated.

  Not surprisingly, the establishment was not happy about these developments. The SOES system was designed to help brokers act on behalf of small, retail investors—mom-and-pop investors around the country calling in their orders—not to enrich a fleet of high-energy, super-motivated day traders pushing the boundaries of speed and arbitrage. (The extent to which the SOES bandits were legitimately representing investors was often a point of contention.) Market makers did everything they could to have them censored. A technological and regulatory arms race proceeded between the SOES traders seeking new advantages and market makers looking to defend their turf. Datek Securities was repeatedly fined. Maschler and his young crew saw themselves as scrappy Davids taking on the old boy network of fat and happy Goliaths who didn’t want the world to change. They didn’t seem to care too much if they needed to bend or even break a few rules in the process. Needless to say, Nasdaq market makers saw it differently.

  Regulations eventually did evolve in response to these trading strategies, but the technology genie was out of the bottle on Wall Street. Trading was becoming faster, more automated, more democratized, and more accessible, and information more transparent. Perhaps it was inevitable. After all, information wants to be free—to borrow a phrase from the emerging world of computer hacker culture that Levine represented (a saying that, incidentally, goes back to one of the fathers of the personal computing movement, Stewart Brand, who had been on that bus with Ken Kesey).

  Maschler would eventually cash out of Datek Securities and walk away with millions, but he would also incur one of the largest fines in history and be personally banned from the securities industry for life by the SEC. Jeff Citron would go on to work with Levine on his new ventures before leaving Wall Street altogether (he was
banned along with Maschler) to help found the telecommunications company Vonage. But the part of this story that was destined to intersect again with mine and with Nasdaq’s was Levine’s.

  In the years following our first meeting, this young savant would parlay his initial efforts at automated trading into a new ambition—building a virtual marketplace where buyers and sellers could come together without the need for a middleman. He envisioned an electronic exchange that would match Nasdaq trades efficiently, immediately, and transparently. It was the culmination of his dream to build a platform where every trader in the market was on equal footing, whether they were in Milwaukee or Manhattan—a sort of protected, virtual “Island.” He built Island ECN using Linux OS, the open-source UNIX operating system, running on Dell boxes, in another basement, this time at his company headquarters on Broad Street in Manhattan. The servers were set atop wooden pallets to protect them from flooding (another reason for Island’s name).

  Many of the world’s great technical breakthroughs seem simple and even inevitable—in retrospect. That’s part of the illusion created by innovation. Our backward-looking narratives so easily fill in the blanks. All the pieces were already present, we might think to ourselves. It was a natural evolution of the market. Yes, that may be true, but only with the benefit of hindsight. It takes genius to put all the pieces together in a new way. Take Steve Jobs and the iPod. Or Tim Berners-Lee and the World Wide Web. Maybe someone was going to create such breakthroughs, sooner or later. But no one had, until them.

  In his book Where Good Ideas Come From, author Steven Johnson posits that creative genius is less about dramatic, singular leaps of invention and more a creative fusion of many smaller ideas put together in novel ways. Levine’s Island ECN was like that. Trading automation, distributed computing power, near-universal online access, a decade-long bull market, cheaper Intel-based hardware that could compete with mainframes, a version of UNIX (Linux) that was open-source—all of it came together in the right creative fusion. Levine devised an ECN that was built around the Linux kernel, so it was reliable. It could also run on distributed Intel servers, so it was cheap and scalable. Its coding was lean, so it ran fast. Levine understood the market structure down to its fundamentals, so his programming was elegant. And he understood trading, so Island’s feature set was built for the new wave of traders entering the market. Was it an inevitable evolution of trading technology? Maybe. But no one else did it quite so brilliantly.