by Liam Vaughan
Meanwhile, over in Zurich, Garcia’s star was rising. After promising to raise substantial funds, Garcia was invited to join the board of the Swiss arm of Robert F. Kennedy Human Rights, a philanthropic group where he mingled with the city’s business elites. Cruising around town in an eye-catching BMW electric sports car, a Maserati, or a chauffeur-driven limousine, his Russian wife, Ekaterina, in tow, the young entrepreneur cut a striking figure. In November 2013, he was profiled in the prestigious Sunday newspaper Neue Zurcher Zeitung am Sonntag. ‘Switzerland’s most important farmer has an urban workplace: Jesus Alejandro Garcia Alvarez performs his daily work in an office at Zurich central within the sight of the railway station,’ the piece began. ‘In his mid thirties in a dark suit, he differs from the usual Swiss professionals in many respects. While farms in Switzerland have an average size of 22.8 hectares, Mr. Garcia and his family have more than 50,000 hectares on which 60,000 cattle graze.’
In twelve months, IXE had apparently morphed from an advisory and trading business into an agricultural powerhouse with holdings in potatoes, avocados and livestock, and a client roster that included Walmart and US Foods. The article and accompanying photograph painted Garcia in an almost mythic light: the pioneering young heir of a great Latin American dynasty, forced out of Mexico by the constant threat of kidnapping, and now driven to build an empire that would make the father he shadowed as a child proud. Reflecting this shift, the company’s website was completely overhauled. Gone were the references to advisory services for rich people. IXE now described itself as ‘a group of companies worldwide … clustered around our core competencies in agribusiness, commodity trading and venture capital’. An org chart showed nineteen IXE subsidiaries spanning Europe, Asia and the Americas.
With his newfound status, Garcia was invited onto Al Jazeera, CNBC and Bloomberg TV, where he opined monotonously on subjects ranging from food standards to the wonders of quinoa to his latest obsession, Bolivian lithium, which he described as ‘white gold’. IXE didn’t publish any financial accounts, but one publication estimated Garcia’s worth at $280 million, placing him firmly in its list of Switzerland’s three hundred richest people. Another suggested he was a billionaire. ‘Our fathers’ businesses and deals were based on a handshake,’ Garcia said in an interview. ‘This spirit lives on in IXE. Our business relationships are built on trust and a word still counts.’
CHAPTER 17
MR X
Navinder Sarao’s arrest would ultimately involve an army of people from agencies including the CFTC, the FBI, the Metropolitan Police and the Department of Justice, but the investigation started with a lone individual with no affiliation to the government. He was a day trader, like Sarao, grinding out a living in a small prop firm in Chicago who, in 2012, happened to be back-testing his system using data from the day of the Flash Crash when he spotted something the whole world had missed. His identity has never been made public. We’ll call him Mr X.
Mr X is a few years older than Sarao and started his career as part of a pit trading firm’s first forays into electronic trading. For a long time he was at the bottom rung; he struggled to cover the rent. He contemplated whether to pursue a career in his primary love, design. Money was never his major priority, but he was good with computers and over time he got sucked into trading. ‘I liked the problem-solving aspect, looking at markets like a puzzle,’ he says. Eventually, he struck out on his own, renting a desk at various firms before teaming up with a handful of traders and programmers he’d met along the way to start their own outfit. In the beginning, Mr X was a ‘point-and-click’ trader like Sarao, placing every order with his mouse and keyboard. But as algorithms proliferated, he realised he would need to adapt. Working with a programmer from Europe he’d been recommended, he started to develop his own algos – basic ones at first, intended to cut down on keystrokes; then more complex iterations that capitalised on the patterns and relationships he’d observed. ‘I’d describe it as grey box now,’ he says. ‘There’s a lot of automation, but it still takes manual inputs and interaction.’
Mr X, who speaks quickly and continuously, as though he’s worried if he stops he might not get another chance, describes himself as a market maker, a timeworn strategy that involves constantly posting orders on both sides of the market and making money on the bid-ask spread. By providing liquidity, he believes he is offering a service to the wider financial ecosystem as well as earning a decent living. At its core, market making is straightforward – you post a bid to buy an e-mini for, say, $1 and an offer to sell one for $1.02 and, when both orders have been lifted, you pocket two cents. Do that hundreds of thousands of times a day and you have a business. But markets move around quickly, and providing two-way prices can be risky. Making consistent profits irrespective of market conditions is a specialised and complex undertaking. To maximise his returns, Mr X asked his programmer to develop a program that would help him identify his competitors based on the footprints participants leave behind when they trade. The ladder is full of potential clues: the size of one’s orders, how long they’re left for, how far away they are from the current price, and so on. Using his design background, Mr X created tools that allowed him to visualise what was happening in the markets more clearly using things like colour-coding, flashing and charting. ‘You’re trying to figure out what kind of drives the markets, and hopefully, get a sense of other people’s strategies,’ he explains. ‘It can be difficult and speculative, but sometimes patterns emerge.’
In August 2012, at the tail end of a heat wave, Mr X and his developer were updating their detection software to incorporate some new features. With little forethought, they decided to test the upgrade on the most eventful day in recent market history: 6 May 2010. Like everyone else in his industry, Mr X had taken a keen interest in the Flash Crash. He was at his desk when the bottom fell out of the market, and, while he escaped without losses, he was left with a residual feeling of disquiet about the precariousness of the system. In the months that followed, he’d read and reread the SEC/CFTC report, and he was familiar with the work of Kirilenko and his peers. He had some theories of his own, but the decision to run his software over the day of the crash was, at least consciously, the result of blind luck.
When the results came in, two things struck Mr X. The first was the glaring imbalance between resting buy and sell orders in the e-mini order book that day – anomalous even after factoring in all the negative economic sentiment. The second, driving the first, was the monolithic slabs of sell orders beginning three levels above the best offer that started appearing in the ladder at 9.20 a.m. Chicago time and moved in lockstep with the e-mini as it made its descent. For the first couple of hours, the blocks showed up for a few minutes at a time before disappearing again, but from a little after 11 a.m. until 1.40 p.m. – a minute before the market tanked – they were left to stand there like an unassailable wall.
Two years had passed since the government’s inquiry into the causes of the crash, and Mr X had watched as spoofing had risen to the forefront of the trading fraternity’s collective consciousness. He’d read about how nine stock traders and their bosses from New York brokerage Trillium were fined $2.3 million for systematically entering ‘layered, non bona fide’ orders to ‘create the false appearance of market activity’. Trading forums were suddenly awash with, on the one hand, complaints about the preponderance of spoofing, and on the one other, fretful discussions about what was and wasn’t allowed under the new regime. As Mr X examined the e-mini data from 6 May 2010, he was unequivocal: ‘I was immediately sure what I discovered was massive manipulation.’
Mr X checked and rechecked the data, and had drawn-out conversations with his colleagues about conceivable alternative explanations. In the end, though, nothing aside from spoofing made sense. One of the factors that convinced him was the way the seller’s orders reacted when fresh information entered the marketplace. ‘This entity designed some questionable behaviours into its strategy,’ he explains. Many legitim
ate strategies involve loading the book with orders at multiple price levels, which are then cancelled and reposted frequently with the goal of improving the odds of getting to the front of the queue. This algo did the opposite. The orders all moved to the back of the line every time the price moved, greatly reducing the chances they’d ever actually be hit. ‘If this was a valid strategy, the participant would have immediately overhauled the design because he wasn’t getting any fills,’ Mr X says.
Mr X spent the next couple of days in a state of confusion. Was it really possible that one of the biggest intraday crashes in the history of financial markets was at least partially the result of a single entity’s manipulation? And if so, why had nobody else noticed it? It was surreal. Mr X went back to the CFTC/SEC’s findings and read them line by line. A preliminary version of the report states: ‘At or about 1.30 p.m. CET, the electronic limit order book in the e-mini S&P 500 futures market exhibited a significant imbalance of sell orders and buy orders. In the backdrop of declining prices, this imbalance appears to have contributed to a sudden liquidity dislocation despite increased trading volume.’ The agencies, then, had identified the massive preponderance of resting sell orders but had seemingly not probed what – or who – lay behind it. They focused instead on consummated trades, which led them to Waddell & Reed’s door. ‘It’s unfortunate that regulators did not use a more expansive data set,’ says Mr X diplomatically. More perplexing to him is the failure of the CME, whose $2 billion of profits in 2011 could have funded the CFTC for a decade, to identify what was happening in its market. ‘There is no way this behaviour could escape exchange oversight, simply because it was so large,’ he says.
As Mr X considered his next move, he recalled a recent article in the Wall Street Journal by Scott Patterson and Jenny Strasburg titled ‘For Superfast Stock Traders, a Way to Jump Ahead in Line’. The piece told the story of Haim Bodek, the forty-one-year-old founder of an electronic trading firm who had inadvertently stumbled upon what he perceived as evidence of collusive practices between the exchanges and their HFT clients. In December 2009, Bodek was standing at the bar at an industry party when he complained to an executive from one of the newer stock exchanges, Direct Edge, about a recent dip in profitability. The executive quickly diagnosed the problem. Bodek was still placing trades using standard limit orders instead of one of the newfangled order types his exchange and others had quietly created at the request of the HFTs. The executive walked through one particular order type called ‘Hide Not Slide’ that allowed users to conceal their bids and offers until they were consummated and, under certain circumstances, to jump to the front of the queue – the holy grail of many HFT strategies. Bodek was blown away. He scrawled the words ‘HIDE NOT SLIDE’ on a napkin and left the party. ‘Man I feel like an idiot,’ he wrote in a follow-up email a few days later. ‘Never grasped the full negative alpha in a normal day limit.’
But Bodek wasn’t an idiot. He was a wunderkind in the world of electronic trading, the son of a renowned particle physicist who, when he was in his mid-twenties, worked for a firm called Hull Trading that was acquired by Goldman Sachs for half a billion dollars. Bodek had held senior positions at Goldman and UBS, reaching the top of the Wall Street totem pole before packing it in and starting his own firm, Trading Machines, in 2007. He prided himself on his knowledge of the inner workings of the markets, yet he’d only stumbled upon the order-type trick by buying the right guy a drink. What hope did anyone else have? Bodek started using Hide Not Slide orders and his profits improved, but he was still perturbed. Part of what appealed to him about trading in the first place was its Darwinian nature. Sure, some participants were faster or better informed than others, but there was nothing to stop people from investing in better technology. This was something else. By developing enhanced functions and then letting only a small proportion of their customers in on the secret – those who traded the highest volumes, who generated the most commissions and whose executives sat on their committees – the exchanges were stacking the deck. It was as though they had created a cheat code for a computer game and only passed it along to their friends.
Bodek could have kept the code to himself, but instead he chose to take his findings to the SEC, which launched an investigation. Blowing the whistle can be hazardous for one’s career, but Bodek took the view that if nobody came forward, nothing would ever change. The photograph accompanying the Journal article showed him staring into the camera, hands clasped on the hilt of a sword, the sun casting a bright light on his perfectly bald head. While alerting the authorities was a brave move, Bodek’s motives weren’t solely altruistic. A few months earlier, as part of Dodd-Frank, the government had introduced a new whistle-blower programme that provided monetary awards to anyone who came forward with original information that resulted in a successful enforcement action. Bodek, who eschewed his right to anonymity, now stood to receive between 10 per cent and 30 per cent of any penalties the SEC collected.
The article included a quote from Bodek’s lawyer, Shayne Stevenson, from the Seattle-based firm Hagens Berman Sobol Shapiro. Sensing an opportunity, Mr X gave him a call. Stevenson was an idealistic young attorney who’d grown up in a trailer park and set up a workers’ rights organisation at Yale Law School before landing a coveted spot clerking at the US District Court for the Southern District of New York. Smooth talking, with a shaved head and goatee, he could have walked into any white-shoe firm but instead returned to his native state of Washington. ‘Anyone who knows me knows that world isn’t me,’ he says. ‘I grew up in the Pacific Northwest and I love it here. I was never going to be a corporate lawyer. My dad is an environmental biologist. From my office I can see three bodies of water; there’s snowcapped mountains.’
Stevenson worked as a state prosecutor for a few years before leaving to join Hagens Berman as a plaintiff lawyer. A big part of his new job was helping individuals and groups bring actions under the False Claims Act, a piece of legislation that allows for members of the public to blow the whistle against entities that are defrauding the government in exchange for a financial reward. So when the SEC and CFTC whistle-blower programmes were drafted using the same language, he positioned himself as a go-to lawyer in the emerging space. Bodek was his first Dodd-Frank whistle-blower client. ‘Me and Haim actually bonded over our love of straight-edge, punk, thrash and hard-core music, specifically the sweet spot of 1984 to 1989,’ he says. ‘Metallica, Slayer, Anthrax, Dead Kennedys, Minor Threat. We both played in punk bands, and if we weren’t talking about the case we’d be discussing some new Viking metal band.’
When Mr X called and told Stevenson that he’d uncovered evidence of manipulation on the day of the Flash Crash, the lawyer was sceptical. For one thing, he estimates he takes on about one case for every fifty inquiries he receives. ‘You don’t know how many times I’ve been told the S&P 500 is rigged,’ he laughs. For another, most whistle-blowers have some kind of inside information; they’re employees or contractors who have witnessed malfeasance firsthand. What Mr X claimed to have done was more remarkable. If what he was saying was true, he’d taken publicly available data from one of the most closely studied days in the history of the markets and seen something everyone else had overlooked. He might as well be claiming to have spotted a second shooter on the grassy knoll in the grainy footage of John F. Kennedy’s assassination. Yet everything else about Mr X engendered confidence: his two decades of trading experience, his technical expertise, the logical, exacting way he spoke. By the time Mr X had walked Stevenson through the data broken down into charts, images and slow-motion videos, the lawyer was sold. ‘When you see it all laid out in that format with that level of sophistication, there really is no other explanation,’ he says.
With Stevenson’s help, Mr X filled out the CFTC’s whistle-blowing documentation and, in November 2012, sent it to Washington along with some supporting data. While he waited to hear back, he daydreamed about who the mystery entity might be and what a huge controversy it would cau
se if Goldman Sachs or Citadel were found to have caused the Flash Crash. ‘I got pretty obsessed,’ he says. ‘Based on the data I guessed it had to be the work of a large prop trading firm, maybe with an internal clearing arm to shield it from the authorities. The scale and the audacity of the behaviour were so massive I couldn’t imagine it was one individual. As it turned out I was very wrong.’
CHAPTER 18
NAVSAR
The CFTC’s Division of Enforcement is made up of prosecutors, who gather evidence and make cases, and investigators, who usually come from the finance industry and have a greater degree of technical expertise. Historically, the lawyers, who are better paid and come from fancy law schools, ruled the roost, but as markets grew in complexity, they increasingly came to rely on investigators with a grounding in data analysis, trading and maths to make sense of the world. In early 2013, Jessica Harris, a thirty-three-year-old investigator, was bashing out code in her small, dimly lit office in Washington, DC, skate shoes up on the desk, keyboard resting on her knees, when she received a call drafting her onto what would turn out to be the biggest case of her career.
Mr X’s tip-off had landed at the agency a couple of months earlier, and the head of the division handed the assignment to a small satellite office in Kansas City that, coincidentally, was walking distance from the headquarters of Waddell & Reed. The investigation there was being run by a team of Midwestern attorneys – Jo Mettenburg, Jenny Chapin and senior prosecutor Charles ‘Chuck’ Marvine – who had distinguished themselves over the past couple of years by busting a string of small-time Ponzi schemers and scam artists. They were experienced prosecutors, but they had scant knowledge of the high-tech world of electronic trading and, since the initial tip-off, the case had only grown in complexity. Within a few weeks, Mr X had passed along a second batch of data tying the as-yet-unidentified trader to suspected spoofing in the days before and after 6 May, 2010. After that, he produced evidence that suggested the same entity was augmenting its layering of the order book by manually placing and canceling orders in clips of 287 and 189 lots. This spigot of new material wouldn’t stop pumping for over a year.
