Flash Crash

by Liam Vaughan

  ‘HELLO NAD,’ Thakkar wrote to Sarao on 3 October 2011, after mishearing his voicemail. ‘I received your message regarding TT related programming. This is what we specialise in. How can I help you?’ A little over a week later, after they’d spoken on the phone, Nav sent Thakkar an email laying out the functions of the system he wanted Edge to build for him. It read:

  1)*JOIN* — This function will put pending orders along either the bid or offer which will become active once the market has flipped price. So if I placed a 300 lot JOIN on the bid at 51 and it is trading 53—54. If either the 53 Bid or the 54 Offer is traded out then my 300 will appear at the same time at 51 Bid.

  2)*JOIN SIDE* — The same as above but the pending order is only activated if the side it is on is flipped. So, in the above example the 300 lot will only join the 51 Bid if the 54 offer is traded and the price goes 54 Bid. If the price simply goes offered down to 51 Bid then the order is automatically pulled and never activated.

  3)*Back of the Book* — For both of the above order types we need to have the option to keep the order at the back of the book. We achieved this by increasing/decreasing the order by 1 lot after it was activated every time a new order was placed on my JOIN activated order. This started to look a little strange with 1 lots changing all the time, so we will have to make it that the order is increased by 1 every time an order greater than say 20 lots is placed. This value may be subject to change …

  5)*SNAP* — This is a pending order ready to trade as soon as the price is available. For example if the price is trading 47—48 and I want to sell 300 at 49 I place the pending order to do so. As soon as it is bid at 49 my 300 lot to sell will become active, whilst it was hidden up to that point. Vital to this is a box with a volume quantity … the minimum amount that needs to be on the bid for the SNAP order to become active …

  6)*MY ICE — *Similar to SNAP, the difference here being that you are instructing this order to maintain constant snapping of the bid/order until the full order is complete … This order needs to be quick enough to be able to catch all flash orders and spoofers that flash on the bid/offer for a millisecond …

  He signed off:

  Please get back in touch with your thoughts as soon as possible. I have decided to stop trading until I have this application since I am so far behind my competitors so time really is of the essence.

  Regards,

  Nav.

  Nav wanted Edge to make him a program that he could control through an interface that would sit next to the ladder on his screen. It would contain a series of buttons which he could turn on and off with his mouse. The top two order types, JOIN and JOIN SIDE, would be used for spoofs. The idea was that Nav would select how many orders he wished to place, click JOIN or JOIN SIDE, then activate them at a given level. They would remain pending until the next time the e-mini price changed, so as to slip into the book amid the commotion of a price move. To minimise the chances the orders would ever be executed, they would be loaded with the ‘back of the book’ feature that TT had declined to make for Nav two years earlier. When ‘back of the book’ was switched on, it would automatically modify orders by alternately adding and then subtracting a single lot every time a new order arrived, constantly sending them to the back of the queue – like the shopper who leaves their line at the supermarket. On the rare occasions a portion of a spoof order was hit, the remainder would immediately be cancelled.

  While the spoof orders were working their magic, Nav would simultaneously be selling e-minis, building a genuine short position. Seeing all this activity, other participants would also start selling, driving the e-mini lower. Once the price had fallen by a couple of ticks, Nav would surreptitiously start buying back e-minis using the order types he called SNAP and MY ICE, which were designed to remain pending but invisible to other market participants until the requisite number of orders became available, at which point they would strike, viper-like. The goal, as ever, was to give off as little information as possible. Once Nav was flat – after he’d bought back the same number of e-minis he’d sold – he would cancel the spoof orders and wait a few minutes before starting the sequence again.

  As the rest of the industry migrated to fully automated systems, Nav chose to remain in the cockpit, directing all buying and selling himself, only with an enhanced set of weaponry. By the time he approached Edge, his plan was fully developed and he sought no advice on the mechanics of the system or the market. He simply needed someone to help him realise his vision. The language Nav used to describe the program was somewhat idiosyncratic, which made sense; he was an autodidact. Everything he knew, he’d learned by sitting and watching the ladder. With no friends at any HFT firms to guide him and no master’s degree in financial engineering, he’d got to a point where he was able to reverse-engineer the strategies of the ‘nerds’ he renounced and create his own system to beat them. Questions of legality aside, it was a remarkable feat. ‘I would have employed him in a heartbeat,’ says the owner of one HFT firm.

  Thakkar took Nav’s plans and instructed his developers to build a prototype that he sent back on 11 November 2011. He named the file ‘NAVTrader’. Over the next two months, Edge fine-tuned the program and Nav tested it. By late January, they’d made sufficient progress that Nav was willing to sign a contract. Thakkar originally offered to do the work for a fixed price of $12,500, stating that ‘this is sort of below cost for us, but we would do it in hopes to make money by selling it to other customers’. After spending many additional hours getting the program to Nav’s satisfaction, he increased that to $24,200 – still a modest sum considering how much money Nav was making. Before the program was completed, Nav had one request. He didn’t feel comfortable having such a controversial system named after himself, so he asked Thakkar to change the filename from NAVTrader to MASTERCHIEF, the name of the protagonist in the popular video game series Halo. In the games, Master Chief is a soldier who uses state-of-the-art battle armor and sophisticated artificial intelligence to lead humanity to victory in an epic battle against a technologically advanced alien race. Thakkar didn’t honour the request, and the NAVTrader moniker stuck.

  With the government’s clampdown on spoofing, Nav was careful never to explicitly talk about how he planned to use the program, and Thakkar never asked. The programmer made a point not to pry into his clients’ trading strategies, and he would later tell investigators there were potentially legitimate uses for all the functions Edge built. He was certainly well positioned to understand the issues at hand. In spring 2012, Thakkar was one of two dozen or so experts from across the futures industry invited to join a new committee set up by the CFTC to examine how regulators might better understand and oversee high-frequency trading. The subcommittee Thakkar was a part of, Market Microstructure, was tasked with exploring the negative and positive ways automated trading impacted markets, and its members talked about spoofing during their regular conference calls and meetings. (Thakkar would remain on the committee for two years, during which Edge continued working on different iterations of NAVTrader.)

  In summer 2012, with his new weapon ready to unleash, Nav finally found a new broker to replace MF Global, an independent Chicago firm with Irish roots called R. J. O’Brien. His contact there was a forty-two-year-old former MF Global employee named James Prince, who lived in a grand house in the English countryside and would prove as incurious about his client’s tactics as the brokers at GNI. R. J. O’Brien wasn’t prepared to extend Nav the same level of credit as MF Global, though, and the size of his positions came down. The heady days of bulldozing the market around with $200 million spoofs were over. But the loss of firepower was balanced out to some degree by the heightened precision and efficiency that NAVTrader introduced. One day a few months later, Nav used his new weapon for a little over a minute to make $55,000. Another time, he made $23,000 in one hundred seconds. For the Master Chief, milking markets had never been quicker or easier.

  CHAPTER 16

  JESUS ENTERS


  Nav may have been preternaturally gifted at making money, but he had little expertise in how to invest or manage it. Nav erected an impenetrable wall between his home life and his business affairs, and when Daljit, his mother, left the house each morning to work behind the till at a local chemist, or Nachhattar, his father, walked to the doctor’s office to pick up a prescription, they did so with no inkling that their son was a multimillionaire. Money for Nav was an abstraction, and the more of it that flooded in, the more he turned to MacKinnon and Dupont for guidance. By now the pair had taken to describing themselves as Nav’s ‘family office’ – a kind of one-stop shop that handles every aspect of a high-net-worth individual’s financial and investment affairs – and not long after the wind deal was finalised, they introduced him to another opportunity, this one involving a mysterious young Mexican businessman named Jesus.

  Jesus Alejandro Garcia Alvarez was the owner and CEO of a Zurich-based company called IXE, an Aztec word apparently meaning ‘one who shows their face and keeps their word’. Dupont first encountered him in the summer of 2011 when he was in the audience of a presentation Garcia was giving to a roomful of financial advisers, solicitors and accountants in Mayfair. On its website, IXE described itself as a kind of broad-based advisory business, offering everything from asset management and legal advice to ‘ultra sumptuous travel’. That day, though, Garcia was touting a different opportunity. Garcia had moved to Europe five years earlier at the age of thirty, he told the attendees, to grow the family business. The Garcias had high-level relationships around the globe, he said, and had identified an opportunity to capitalise by providing short-term credit to facilitate the trading of ‘physical’ commodities such as coal and cooking oil. The mechanics were somewhat opaque, but the idea was that IXE would act as a middleman between buyers and sellers in exchange for a fee. Rather than waiting months to receive payment for a shipment of coal, for example, a supplier in China would accept a reduced sum from Garcia’s firm to get the money now. IXE would deal only with governments and companies with the best reputations; and it would protect itself by making sure the transacting parties signed legally binding ‘letters of credit’ in which they pledged to abide by the terms of the deal. IXE was already making big profits, Garcia said, but there was so much demand for capital it had decided to open up the opportunity to outside investors, which is where the introducers came in.

  ‘We are offering alternative investment vehicles that provide constant returns to investors,’ Garcia explained in an accent so thick it left the audience straining to understand him. ‘The investment in real economy makes the advantages obvious – investors are benefitting from constant returns generated from actual transactions with zero speculation and zero volatility.’

  Garcia didn’t exude gravitas. He was small and stocky and thickset with jet-black eyes that blinked perpetually and a mouth that turned downward when he spoke, like a ventriloquist dummy. He delivered his presentation in a monotone that veered in and out of coherence. Bizarrely, he was accompanied at the meeting by a smartly dressed British couple in their sixties named Lynn Adamson and Chris Sawicki, who had agreed to be IXE’s UK agents despite meeting Garcia themselves only a short while earlier. Adamson, who sported a cravat and drove a sports car, had spent most of her career advising middle managers on how to reduce stress, but she wound up writing IXE’s pitch documents because Garcia’s English wasn’t up to snuff.

  Garcia’s words may have been faltering, but his pitch was undeniably enticing. IXE was offering ‘participants’ (it eschewed the word ‘investors’) who made a minimum deposit of $1 million a guaranteed return of 8 per cent per year. The money would be held in their own personal accounts with a Swiss bank called Hinduja and could be accessed only with their sign-off. Since the cash would exclusively be used as a ‘backstop’ in pre-agreed trades, it would never actually be put at risk. Introducers, such as MacKinnon and Dupont, would get a meaty 4.5 per cent a year on everything their clients invested. Asked about IXE’s pedigree, Garcia told the attendees it was a ‘sister company’ of a $300 billion Dubai-based entity called ETA Star, ‘the largest conglomerate in the Middle East’. IXE also claimed to be regulated by FINMA, the Swiss financial regulator.

  In the weeks that followed, MacKinnon and Dupont gently nudged Nav towards the IXE opportunity, and in July 2012 they accompanied him to Zurich to meet Garcia in person. It was Nav’s first trip out of the country since visiting relatives in India as a boy, and when Dupont picked him up on the way to Heathrow in an Aston Martin – Nav asked him not to park on his street in case anyone saw them – the mood was jubilant. Their guide in Switzerland was Adamson’s partner, Sawicki, who drove them to a nondescript building with a creaky lift on a residential street where they were surprised to discover the Swiss bank Hinduja was based. After obtaining written confirmation from one of the bank’s directors that Nav’s money couldn’t be touched without his approval, the party proceeded to IXE’s headquarters in the centre of town.

  IXE’s office was small but tastefully decorated, with wood panelling and a door thick enough for a bank vault, creating a sense of theatre. In a boardroom overlooking the Limmat River, Garcia talked Nav through what he called his ‘Physical Commodity Participation’ opportunity one more time. The meeting was something of a slog, as Garcia’s impenetrable accent butted up against Nav’s inner-city English, but Nav was enthusiastic about the concept and excited about the returns on offer. Afterwards, they went for lunch by the river, where Garcia and Nav bonded over football while MacKinnon and Dupont gritted their teeth and Sawicki sweated through his suit. By the time the bill arrived, Nav had made up his mind to participate.

  Around the same time, however, a development back in London threatened to derail the deal before it got off the ground. One of the introducers at Garcia’s Mayfair presentation had convinced a sovereign wealth fund to invest hundreds of millions of pounds in the venture. Before signing contracts, the fund had hired a corporate investigator to carry out some due diligence, and, based on what it discovered, it had decided to walk away. The investigator had found a 2010 legal complaint filed in Florida in which both Garcia and IXE were accused by Mongolia’s central bank of participating in an elaborate scheme to rob it of more than $20 million. The alleged scam, which was based around letters of credit, was the brainchild of a Florida native named Burton Greenberg, who would later serve an eight-year stretch for an unrelated fraud.

  When Garcia discovered that an introducer was spreading potentially damaging information about him, he instructed a lawyer to draft a rebuttal vigorously denying the allegations and pointing out that neither he nor IXE was ever a defendant in the case.[fn1] Nav was never told about the Mongolian suit, and after negotiating with Garcia to bump up his interest to 9 per cent a year, he signed off on the transfer of $17 million from his offshore company to a new account in his name at Hinduja Bank. Three months later, the first statement arrived. Where previously Nav had been lucky to receive a couple of thousand dollars a quarter in interest, he was now banking in the region of $380,000. Via the power of compounding interest alone, his funds stood to double within eight years. Nav was so intoxicated with the influx that he suggested transferring the $15 million or so sitting in Cranwood, the wind holding company he co-owned, into Hinduja as well, so it could generate some decent returns until it was required.

  IXE posed something of a quandary for MacKinnon and Dupont. Thanks to Nav’s initial deposit, they were now receiving a residual $500,000 a year in commissions, and they didn’t have to do a thing. They had reservations about Garcia and the people he surrounded himself with; but they reasoned it was ultimately up to Nav what he did with his money. Cranwood was a different matter. As directors in Wind Energy Scotland, many of their day-to-day expenses were covered by the Isle of Man entity; and if Martin Davie, its head, achieved half of what he said he could, they were set to become very wealthy. Preserving that capital was paramount. On the other hand, Hinduja seemed to be a respectable, inde
pendent entity, and if they did deposit the additional $15 million with IXE, the pair would receive an extra $450,000 a year in commissions. In the end, they agreed: Cranwood’s money would be transferred to a second account at Hinduja, from which Wind Energy Scotland’s costs would be doled out on a rolling quarterly basis, as needed.

  Shortly after Nav made his first deposit, Western governments announced sanctions against Iran, and IXE wrote to investors telling them they would need to move their money out of Hinduja, which had Iranian connections, to Morgan Stanley, where IXE also held an account. It was an unwelcome development, but IXE provided written assurances that no money could be accessed without account holders’ approval. Comforted by Morgan Stanley’s involvement, Nav approved the transfer.

  By the following summer, Nav had made two further deposits into his IXE accounts, bringing his total investment to around $50 million. Like the George Eliot character Silas Marner hoarding his gold coins, Nav watched the numbers in his bank accounts grow, never withdrawing a penny for himself or his family. He liked to boast he could buy a Big Mac every minute with the interest he was making. Meanwhile, those around him enjoyed the spoils of their success. MacKinnon and Dupont, who were now receiving commissions of around $750,000 a year each from IXE alone, moved from the outskirts of Mayfair to its epicentre, Berkeley Square, renting an office in a plush town house, which they decorated with a giant map of Scotland and an antique rifle. Away from work, MacKinnon remodelled his farmhouse, which had a tennis court, a swimming pool and a wine cellar. Dupont bought an elegant apartment in the country and a Ferrari California. Every quarter, money rolled in via a company set up by the husband-and-wife duo of Adamson and Sawicki, who, after a lifetime of near misses and failed ventures themselves, had finally struck gold.