On May 6th, 2010, the evolving European debt crisis was already taking its toll on the morning’s markets. Into this shaky market, the traditional explanation goes, a large fundamental trader, using a poorly-understood computer algorithm, set in motion the Flash Crash by selling a very significant dollar volume of S&P 500 futures over the course of twenty minutes. Yet recent news alleges that the manipulative activity of a single trader, not this large futures order, was a contributing factor of the crash.
Neither a large order nor manipulative activity should have been able to cause the Crash. If one reasonably small-time bad guy can cause a crash in the market, the problem is not the bad guy, it’s the market structure. Our modern markets have become a fragmented and interactively complex system that is impossible to predict or understand. Decades of research shows that systems with this level of complexity, especially with limited slack to correct errors, are not merely vulnerable, but prone, to catastrophe.
On the day of the Crash, since many firms simultaneously participate in the closely linked futures and equities markets, the futures activity disrupted the equity markets. Further, trading firms’ use of similar risk models, delays created by exchanges in matching trades and disseminating market data, and the uncertainty as to whether equity trades would stand or be broken all contributed to the market’s chaos. In dynamic markets, errors can escalate so quickly that operators cannot understand or respond to worsening conditions.
Following the Crash, the SEC and CFTC issued a joint report that identified multiple causal factors and the benefit of preventative measures, like the Chicago Mercantile Exchange’s automatic trading halt, that allowed the market to briefly pause and recover. Since then, the SEC has taken limited steps, including implementing single-stock circuit breakers and reducing ambiguities regarding erroneous trades. This is a pattern we see all too often in the wake of catastrophic events: a prescriptive solution is quickly deployed, fixing the proximate technological causes of a failure, but the roots of the failure persist: in this case, the complexity of the markets.
One of the most vexing features of the Flash Crash is how its volatility emerged from the market’s otherwise normal interactions (even if the recently alleged manipulation was present on May 6th, it was also present on hundreds of other trading days). Rather, the Crash’s root causes stem from layers of technology, interconnections between exchanges and brokers, complex order types that confound even sophisticated traders and exchanges themselves, and the burdens and requirements imposed by rules that continue to emerge from multiple independent regulatory groups.
Worse still, even rules that were once easy to implement (like marking sales short) have become nearly impossible to follow due to interactions between computers, strategies, number of trading venues, and the speed at which information travels.
In the five years following the Flash Crash, we have reaped the consequences of this complexity in numerous mishaps by clearly well-intended, major market participants. Though these failures have varied in character and were caused by seemingly distinct causes, they are all rooted in the inability to manage the technological complexity of the current market system. These failures have occurred at exchanges, like NASDAQ’s handling of the Facebook IPO or the CME’s accidental release of 30,000 test orders into live markets; in obscure pieces of market infrastructure like the hours-long interruption in NASDAQ’s Securities Information Processor, which halted trading market-wide in NASDAQ-listed stocks; and at major market participants, like Knight Capital’s $440 million trading “glitch.”
These root causes are intrinsic to our current market structure. They will continue to cause unpredictable failures. How can regulators go from individual technical changes and punishments to insights that could actually reduce the potential for catastrophic failure?
First, regulators need to move toward radical simplification of our market structure. Each new order type that an exchange adopts, each new pilot program, and each new regulatory interpretation may appear beneficial at the margin. However, in aggregate, these requirements result in an increasingly costly, complex, and error-prone market. This creates an environment for failures, the very kind of consequential errors that regulators should be most concerned with preventing. This move toward simplicity should begin with a reexamination of the complexity that is created by Reg NMS, the cornerstone of the equity market structure.
Second, regulators need to move toward a culture that encourages collective learning. The errors that most threaten the fair and orderly operation of our nation’s markets are those that stem from complexity. These errors get worse, not better, when a punitive culture exists. Bad actors–like market manipulators–should be punished. But regulators should undertake investigatory responses to technical failures without resorting to enforcement actions–much like their initial response to the Flash Crash. They should also provide firms a mechanism for the confidential and anonymous sharing of information about technology glitches and near-misses. In the current environment, firms suppress lessons learned from such technological missteps, even internally. As a result, the industry is unable to learn from its collective experience. Enforcement actions, and their implication of criminality, further undermine the potential for learning and the stability of the markets.
Moving forward with reforms is paramount. Regulators should not delay. We have already seen numerous errors similar to the Flash Crash. The complexity and interconnectedness of the markets makes such failures inevitable. Unless regulators themselves stop requiring such complexity, we should expect this trend of destabilizing errors to increase to the point where they are a normal feature of one of our most important national assets.
Thanks to Steven Lofchie, a partner at Cadwalader, Wickersham & Taft LLP and the Co-Chair of the Financial Services Group, and James Owen Weatherall an Associate Professor at the University of California, Irvine, and author of The Physics of Wall Street, for help preparing this article.
This piece originally ran in Forbes on May 6th, 2015.