Flash Crash

by Liam Vaughan

  As the Flash Crash had demonstrated, regulators weren’t just unable to monitor in real time what was going on in their markets. Even seven months on, the SEC didn’t have the data it would have required to understand what happened to equity markets over a single half-hour period at a granular level. If the umpires couldn’t watch the game, how could they hope to referee it? To remedy the issue, Schapiro proposed something called the Consolidated Audit Trail, a vast data repository that would allow regulators to track orders across the market and identify the brokers handling them. It would take at least four years to build and cost an estimated $4 billion. In the meantime, the financial cops would have to continue relying on the private exchanges to alert them to problems.

  As financial markets had evolved, the government’s ability to oversee them had fallen further and further behind. Fundamentally, it was a question of resources. The SEC’s budget in 2010 was just over a billion dollars, while the CFTC got by on less than $200 million. The year before, meanwhile, HFT giant Citadel’s founder, Ken Griffin, personally took home $900 million. Each year the agencies begged for more money to hire experts and upgrade their aging technology and each year they were knocked back. There was also a glaring skills gap. At the CFTC, where lawyers reigned, investigators and economists with up-to-date computer skills and markets experience were in short supply, and it was easy to understand why. Even the most civic-minded young programmer would struggle to turn down hundreds of thousands of dollars on Wall Street or in Chicago for a $70,000 base salary, no sleep and bad coffee at one of the agencies. ‘Resources are a significant concern,’ said Schapiro with practised restraint. ‘We’re trying to bring in new skill sets, people with experience in algorithmic trading … on trading desks and hedge funds … to try to help us have the capability to do the job we’ve been charged with doing.’

  About an hour into proceedings, Senator Jack Reed, a former serviceman from Rhode Island, shifted the conversation from the practical to the philosophical: ‘The presumption of most people who own a few stocks is that the value of stocks, the liquidity of stocks is directly a function of their economic value, the same thing with debt instruments and derivatives. One of the issues we have to deal with is, with the proliferation of these algorithmic, high-frequency traders, some of these algorithms don’t take into account the fundamentals of the instrument, the economic value, the dividends, the status of the municipality issuing them. They are simply saying, “if enough of these are sold then we start selling, and then if we start selling another algorithm kicks in”. To the extent that we get further away from the economic values here, does that not only cause concern, but is that something that’s good for the economy? It may be a naïve question but I’ll pose it.’

  ‘It’s not a naïve question at all,’ Schapiro replied. ‘It’s sort of the fundamental question we’re all grappling with.’

  The Flash Crash had laid bare the regulators’ limitations, but more than that, it had awoken the wider population to the fact that the entire structure of the financial markets had shifted under their feet without them noticing. Those few anxious minutes, when it seemed conceivable that the whole edifice would come crashing down, had led to a reckoning. The senators spoke for all Americans when they pondered what the repercussions were of completely automating a system that determined the value of our companies, our savings and the food and resources we consumed; in whose interests it was for securities to change hands thousands of times per minute; who the winners and losers were from this paradigm shift; and what would happen if the technology were to be abused.

  Later in the hearing, after Gensler and Schapiro had left, the senators asked the owner of an HFT firm named Manoj Narang what risks high-frequency trading posed to the markets. ‘It takes capital to move markets … and virtually every HFT out there controls very little capital … so it’s entirely implausible,’ he said in a clipped tone. When another panellist suggested that some HFT firms might deliberately seek to manipulate markets, Narang was defiant: ‘I don’t know of any HFT strategies currently in use or that can be hypothetically conceived of that are destabilising to the market.’ The exchange was symptomatic of a growing divide between supporters and critics of HFT that would play out in universities, cable news studios, trading floors and newspaper columns for years to come. Advocates of HFT argued it brought efficiencies to the marketplace, lowering bid-ask spreads, a widely accepted measure of value for money; and adding liquidity, allowing market participants to trade quickly and easily at stable, transparent prices. They painted their opponents as reactionary has-beens tilting at windmills. The naysayers portrayed HFTs as leeches who made false claims and brought instability. By 2011, the debate was no longer just a theoretical one. For the first time, governments around the world seriously considered clamping down on high-frequency trading. It wasn’t simply a matter of stability; there was also the question of fairness. Was it really appropriate that a small group of lightly regulated firms was creaming more than $10 billion out of the markets each year, and at what point, if ever, should the government intervene? With Dodd-Frank still trundling through the system, agencies mulled whether fresh rules might be incorporated. The CFTC launched a public consultation and invited two dozen traders, brokers, exchange executives and professors to Washington to discuss whether increased regulation was required. The SEC undertook its own review of equity markets and invited parties to submit their thoughts. Similar inquiries took place in Germany, Italy and the UK. After years spent hiding in the shadows, HFT was in the spotlight.

  Proposals for reform varied widely. One of the more contentious was a transaction tax, or ‘Tobin Tax’, which, depending on its size, would either make it uneconomical for HFT firms to place and cancel so many orders, or at least raise some additional funds for the government. Another was bringing in ‘speed bumps’ to slow down orders, eradicating the HFTs’ advantage. Other options included making HFT firms register the code for their strategies, levying fines for cancelling too many orders and forcing market makers to quote prices regardless of how volatile conditions got. ‘We basically viewed it as a public safety issue,’ recalls one former CFTC employee. ‘There are safety measures and speed limits for automobiles. Should there be something similar in financial markets?’

  For the most part, the HFT firms and the exchanges where they operated argued that reform was either unwarranted or could be achieved voluntarily. Fifteen of the biggest firms formed a body called the Principal Traders Group, which commissioned professors to write papers espousing the benefits of HFT, gave speeches on the hysteria of the media and funnelled donations to their political allies, including Chicago’s mayor, Rahm Emanuel, who visited the CFTC to lobby against one proposed rule. The group appointed an ex-CFTC economist to make its case and, in November 2010, took the agency’s chairman, Gensler, out for steaks in Chicago to press their interests en masse in a more informal setting. For once, Gensler, a former senior executive at Goldman Sachs, wasn’t the richest one at the table.

  While the HFT lobby was slick, well funded and organised, their opponents were a more motley crew. Within the CFTC itself, one of the regulator’s five politically appointed commissioners, the Democrat Bart Chilton, pushed for tighter rules and appeared on business television describing HFT firms as ‘cheetahs’. With cowboy boots and blond locks that gave him the appearance of an aging He-Man, he gamely debated the subject on talk shows, but he represented the minority voice. Also invited to sit on various panels were Joseph Saluzzi and Sal Arnuk, two straight-talking Italian-American brokers from New Jersey, who found it galling to be debating the finer points of regulation when they considered it patently obvious that the whole system was rigged against their clients, the pension and savings funds. Their influential 2012 book, Broken Markets, was among the first to highlight what they saw as the inherent iniquities of HFT in equities markets.

  A mysterious commentator writing under the pseudonym R. T. Leuchtkafer – German for ‘firefly’ – captured the att
ention of the industry with a series of erudite and scathing critiques of the modern market structure that were picked up by the business press. Like Daniel Defoe three hundred years earlier, he painted a picture of a system set up for the benefit of the few. ‘It is no argument to say the advantages of HFT firms are available to anyone,’ he wrote in one widely publicised letter to the SEC. ‘They are not. They are only available to those with the capital and regulatory latitude to pay for those advantages … The corruptible but regulated dealer has been replaced by largely unaccountable and unregulated firms. You should not underestimate the widespread and legitimate anger at these firms.’

  Perhaps the loudest member of the resistance was Eric Hunsader, the founder of a company called Nanex, which analysed and repackaged oceans of trading data from an office above a barbershop in Winnetka, Illinois. Hunsader detested HFT and built up a cult following on Twitter by posting charts and videos that he said showed spoofing and other forms of cheating. A few weeks after the Flash Crash, Hunsader and his colleagues were scouring data when they noticed that some exchanges were intermittently being bombarded with huge volumes of orders in short bursts, sometimes as many as five thousand in a second. The activity had little impact on prices, but it momentarily slowed the exchanges, leading Nanex to conclude that unscrupulous players were flooding the system with false signals to confound their rivals and gain an advantage. The phenomenon showed up in strange-looking charts that Nanex likened to crop circles and gave names like ‘crystal triangle’ and ‘chalice’. Intriguingly, one occurred as the e-mini collapsed on 6 May. Nanex dubbed the practice ‘quote stuffing’ and suggested it was a major, unspoken cause of the crash.

  In July 2010, Hunsader was invited to Washington to present his findings to a roomful of people from the CFTC and the SEC, who listened politely and asked questions. But when the report came out, ‘quote stuffing’ barely got a mention. In Hunsader’s mind the reason was clear: it was a whitewash. He persuaded Waddell & Reed to send him its trading records and published his own alternative Flash Crash report that vindicated the Kansas City fund manager and said it had been made a scapegoat for the HFT industry. ‘Why would SEC regulators deny that Quote Stuffing exists and that our analyses are “the stuff of conspiracy theories?”’ he asked in one blog post. Maybe ‘the regulators don’t wish to acknowledge that the playground they created has been taken over’ and admit ‘their role as enablers?’

  As the debate continued into 2012, two events kept algorithmic trading on the agenda. In May, Facebook’s long-awaited stock market listing was dogged by technical problems; and, three months later, a bungled software upgrade at Knight Capital Group caused the then-largest intermediary in US equities to fire off erroneous orders for 150 companies. By the time Knight got a handle on the issue, it had lost $460 million, and it was sold soon afterwards for a fraction of its valuation before the incident. Yet the more time passed after the Flash Crash, the more the white heat of fervour dissipated. HFT got bigger and migrated to new markets, but there was no repeat of 6 May 2010; no multimarket ‘Splash Crash’, as the doom-mongers had predicted. ‘There was a lot about HFT that was contested,’ recalls a senior government official. ‘For every person that said these algos were bad, others would say, “No, they’re good.” Empirical data was hard to come by. We needed to have a degree of conviction that not only did we understand the phenomenon, but that regulation was a good response to the situation. We weren’t going to regulate something just because a bunch of people complained about it.’

  For every proposed reform, the HFT lobby was ready with a counterargument. A transaction tax was too onerous and likely to damage the US economy; speed bumps were impractical from a technical point of view; forcing market makers to stick around in volatile periods would result only in bankruptcies; making firms send their secret sauce recipes to regulators posed unreasonable security risks. Some small changes crept in. The CME and other exchanges, which already tracked what proportion of their customers’ orders were cancelled, started penalising outliers with bigger fines. But the window for a more fundamental overhaul of the market closed. The government’s attentions shifted to the rising threat of cybersecurity, the agencies went back to finishing Dodd-Frank, and the assorted committees set up to examine HFT quietly disbanded. ‘It was strange really, nobody ever said it was finished. We just stopped meeting up,’ recalls one elderly committee member.

  On the fourth floor of the CFTC, one unlikely glimmer of resistance still flickered. After the Flash Crash report, Kirilenko had been promoted to chief economist, and, with little budget at his disposal, he focused on attracting students from the best universities to spend time at the agency. By 2012, there were around forty paid and unpaid researchers devoted to understanding the shifting financial landscape. ‘The old ways of modeling and understanding financial markets just weren’t working anymore, they couldn’t capture these new high-speed, high-volume markets,’ says Steve Yang, a University of Virginia PhD who was among them. ‘Andrei had this grand ambition to build a research enterprise to address that.’

  Kirilenko prided himself on being a dispassionate scientist, but there was something about the industry’s ‘nothing to see here’ attitude that stuck in his craw. The general public and the press may not have been able to grapple with the finer points of market microstructure, but he sure as hell could; and his work on the crash had shown him that some of the biggest HFT firms weren’t just selflessly providing liquidity. It was a convenient illusion. In late 2012, Adam Clark-Joseph, a young Harvard PhD who spent time at the CFTC, issued a paper titled ‘Exploratory Trading’ that posited that HFT firms routinely fired small, loss-making orders into the market like sonar to gather what he described as ‘private’ information about market conditions before placing much larger orders when they knew they were more likely to profit. And in November, Kirilenko and two colleagues put out a paper titled ‘The Trading Profits of High Frequency Traders’. With each publication, the group pulled back the curtain a little further.

  Using granular data on trading in the e-mini, Kirilenko et al. were able to substantiate that a vanguard of HFT firms had found a way to make large and extremely consistent profits without taking significant risks. These elite firms, which the professors didn’t name, had proven remarkably resistant to competition and made most of their money from retail and institutional market participants – in other words, the pension funds, mutual funds and day traders they claimed to be making things better for. On 3 December 2012, the New York Times ran an article with the headline ‘High-Speed Traders Profit at Expense of Ordinary Investors, a Study Says,’ complete with a photograph of a stern-looking Kirilenko. The Wall Street Journal published its own account, which quoted the authors asking whether the ‘arms race for ever-increasing speed and technological sophistication’ had any social value. It was incendiary stuff from a government employee, and the following week a letter arrived from the CME’s lawyers, Skadden, Arps, accusing the CFTC of breaching data confidentiality laws and putting the exchange’s customers’ ‘trade secrets’ at risk. As a regulatory agency, the CFTC was permitted – indeed, mandated – to conduct research. But granting outside academics like Kirilenko’s recruits access to sensitive data for use in non-CFTC papers was beyond the pale, the CME claimed.

  Amid pressure from the HFT lobby, Gensler shut off access to the servers for all noncontracted staff. Two months later, he dismissed twenty-one research economists on temporary contracts. For more than a year, no research was commissioned by the CFTC into high-frequency trading or anything else. Several papers never saw the light of day. ‘This was fundamental research that was essential to the public’s understanding of the way markets worked, and it was flagrantly suppressed,’ says the University of Maryland’s Albert Kyle. The purge took place without the involvement of Kirilenko, who left the CFTC a few days after the Times article ran to take up a prearranged position as a professor at MIT. An inquiry into the CFTC’s handling of the affair later concluded
that while the Office of the Chief Economist’s processes for bringing on researchers and keeping data secure were sloppy, there was never actually any breach of the rules on confidentiality. Kirilenko’s researchers had always been careful to ensure that no individual firms or strategies could be identified in their work, and the CME’s allegations were without foundation. In the end it didn’t matter. A sharp jolt of pressure from the HFT lobby was all it took to yank the curtain closed.

  CHAPTER 14

  THOUGHT CRIME

  The CFTC did introduce one significant change to the way futures markets were policed, and it was something that both pleased the HFT community and placed Sarao in the agency’s crosshairs: an outlaw on spoofing. It was brought in to draw a line under the government’s long, woeful record at stamping out cheating.

  King Jack Sturges. Mr Copper. The Boy Plunger. The Hunt Brothers. Crazy Harper. The annals of futures markets are lit up with the names of financial buccaneers who made headlines and fortunes by pushing prices around while the authorities stood helplessly by. In the late nineteenth century, it was gold; in the 1920s, grain; by the fifties and sixties, onions, coffee and soya beans. The commodities changed, but the methods were consistent: a trader with deep pockets quietly cornered supply and forced prices up; an unscrupulous short-seller spread negative rumours and pushed them lower; supposed competitors formed a nefarious pact. In 1974, after a particularly egregious episode known as the ‘Great Grain Robbery’, President Gerald Ford created the CFTC as a kind of designated cop for futures akin to the SEC. But a few years later, when three brothers from Texas monopolised the silver market and pushed prices up by 713 per cent, it was clear little had changed.