October 31, 2010
Congratulations to the SEC and the CFTC on their joint report "Findings Regarding the Market Events of May 6, 2010" ("Joint Report"), and congratulations to CFTC Senior Financial Economist Andrei Kirilenko and his co-authors for their paper "The Flash Crash: The Impact of High Frequency Trading on an Electronic Market" ("Kirilenko").
The two should be read together to understand the Flash Crash. The Joint Report details how a fratricidal storm among market makers in the futures market quickly spread to the equities market. Kirilenko stands as the authoritative account of what happened in the futures market that day, as well as the most detailed examination of high-frequency market makers yet published.
These works end the debate. High-frequency market maker firms are not "passive liquidity providers," as they long claimed. The phrase is absurd and obsolete. They are very active and aggressive traders, committing fratricide when it suits them, or withdrawing altogether from volatile markets. With inventory half-lives measured in seconds, when these market makers reach their risk thresholds and start liquidating inventory - without regard to time or price - they can easily stoke a self-sustaining firestorm while prices collapse.
The Joint Report and Kirilenko both start with a single, large e-mini SP 500 seller in an already tender market, a seller trading by algorithm, who innocently if perhaps foolishly set off a chain of events that erased $1 trillion in a few minutes by itself trading without regard to time and price.
The CME Group, owner of the Chicago Mercantile Exchange, pointed out that the suspect e-mini order was entirely legitimate, that it came from an institutional asset manager (that is, the public), and was little more than one percent of the e-mini's daily volume and less than nine percent of e-mini volume during and immediately after the crash. Later analysis showed that during the last three and a half minutes of the rout, as new losses reached about $700 billion, the order was less than five percent of total volume. In the last 15 seconds, when losses increased by about $200 billion, the order was less than one percent of total volume.
How did this small bit of total volume cause the Flash Crash? Kirilenko provides detail from the futures market. He divides market participants into three main categories based on inventory or portfolio holding periods - intermediary, fundamental, and opportunistic traders. Fundamental and opportunistic traders have longer-term directional strategies and hold inventory, while intermediary traders don't have long-term directional strategies and don't hold inventory. Kirilenko defines the most active intermediaries as high-frequency traders, which they undoubtedly are.
On May 6, an already bad day in the markets, buyers were thin, and the large e-mini seller, together with other sellers, exhausted the few remaining buyers, including high-frequency traders. High-frequency traders hit their risk limits and turned around and started selling, competing with fundamental sellers for any remaining buy-side liquidity. But the only remaining liquidity came from other high-frequency traders, who themselves quickly hit their limits, cart wheeled, and started selling aggressively. The futures market quickly turned into high-frequency fratricide, a fratricide the Joint Report and Kirilenko both call "hot potato" trading, as high-frequency firms dumped inventory onto one another at lower and lower prices. Firms also fled the market altogether, accelerating the sell-off. In the final moments, the original e-mini seller had largely exited the market, or couldn't keep up with it, but events were already irretrievably out of control.
The same pattern unfolded in the equities markets, accelerated by internalizers who saw the mayhem and passed sell orders off to the exchanges - their own version of "hot potato."
Punch drunk, markets stabilized and rebounded after a short halt in the futures market, when other market participants - Kirilenko identifies them as fundamental and opportunistic buyers - realized what happened and jumped into the market.
Fair enough, some might say. Markets panic, and sometimes for no reason. But in the equities markets the larger HFT firms register as formal market makers, receiving a variety of regulatory advantages, including greater leverage. All of this extends their enormous reach and power to create or sustain panic. In the past, they fulfilled certain obligations and observed certain restraints as a quid pro quo for those advantages, a quid pro quo intended to keep them in the market when markets were under stress and to prevent them from adding to that stress. Over the past few years, however, decades-long obligations and restraints all but disappeared, while many advantages stayed.
Computing power also opened market making to a field of unregistered, or informal, high-frequency market makers, what Paul Kedrosky termed the "shadow liquidity system." Exchanges will pay you to do it, too, just as they pay formal market makers, and require little in return.
The result is a loose confederation of unregulated, or lightly regulated, high-frequency market makers. As a group, they participate in as much as 70 percent of all equity trading, and some individual firms are estimated to participate in as much as 10 to 20 percent of equity trading. They gorge on data feeds containing what many consider confidential order information - a scandal where reserve and hidden orders are revealed to the subscribers of these feeds - they play hot potato in volatile markets, and they widen or pull their quotes and play hide-and-seek if even a single hedge hits an unseen and unknowable tipping point.
They do it because they can. It doesn't happen because of a computer malfunction. It happens because it is designed to happen. It is not a bug, it is a feature.
The flash crash was the most spectacular example of this feature of US markets, but it is not the only example. For years, there have been mini-flash crashes in individual stocks. Usually these episodes earned just a perplexed "oops" from participants and regulators.
In the autumn of 2008, though, with the financial system on the brink, stocks whipsawed, foremost because of fundamentals, but also - importantly - because of the machinations of the shadow liquidity system itself. Volatility reached record highs as high-frequency firms played hot potato and hide-and-seek, day after day, month after month, with the biggest listed companies in the world.
As I wrote in my April 16, 2010 comment, because of all of this we must re-regulate market making. Critics of regulation point out that no regulator can ever force high-frequency firms to stay in the market when a freight train comes charging down the tracks. That is obviously true, but the flash crash was a freight train of their own design and manufacture. These firms lobbied for deregulation and lobbied to keep their many advantages over everyone else. So it is ironic the freight train they ran from on May 6 was just another unregulated high-frequency market maker, one that fired off a thousand orders a microsecond earlier, playing hot potato simply because it can. The flash crash was high-frequency brother against brother, with the public and listed companies shredded in the crossfire. This is the very definition of systemic risk.
Bona fide market making has been deregulated to the point where it is nothing more than scalping, with de minimis quoting obligations and no restraints on when these firms can demand liquidity and shock prices. Many dealers used to have inventory half-lives of at least several days. Kirilenko discovered half-lives of less than two minutes in the futures market. Tradebot reportedly flips its inventory in 11 seconds. "Liquidity providers" with the freedom to provide liquidity for just a few seconds before demanding liquidity without regard to time or price are an unpredictable prelude to inevitable disaster, and our markets now completely depend on them.
Unfortunately, we can't expect competition to solve these problems. Stock exchanges tried to compete against these deregulated market makers, but to survive the exchanges had to adopt the same model. The scalping model is cheaper, and because it is cheaper it can post tighter quotes, and because it posts tighter quotes everyone must route to it. We didn't fret about disaster because we couldn't see or imagine it. The scalper's price effects can be transient - just as the Flash Crash was transient - and often evident only in volatile markets or illiquid stocks, where there are already easy explanations for what we see. Of course, illiquid stocks are frequently early-stage companies, the economy's job creators, but we turned them over to scalpers anyway.
In the meantime, SP 500 spreads shrank and trading costs went down, and whether this happened because of automation or decimalization or deregulation, or some combination, it didn't matter. The high frequency firms and their academic and regulatory co-religionists took all the credit. And so exchanges relaxed or eliminated regulations that market maker quotes had to relate to the market, and pointed to the deregulated exchanges as the reason why. The exchanges had to compete. The competitive solution is to devolve. Scalping became the heart of our equities markets.
As I wrote here weeks before the Flash Crash - you should not underestimate the widespread and legitimate anger at these firms. Please regulate them.
R. T. Leuchtkafer
(Portions first appeared in FT.COM/FT Trading Room.)