Anyone who was sitting at a trading desk on May 6, 2010 -- as I was -- remembers the flash crash as one of the most unusual trading days ever recorded. While on a much smaller scale, yesterday's Twitter-induced mini flash crash, or "flash crash 2.0," is likely to have important ramifications. Not only have structural weaknesses in the nature of trading news been highlighted, but the potential impact of weak security in social networking will be discussed in a new light. The ensuing investigation, which will primarily focus on the hackers who initiated the bogus tweet, may reveal a great deal about how market structures need to be addressed.
Flash crash 1.0
The first flash crash was determined to have been caused by the sale of a large block of S&P 500 (SNPINDEX:^GSPC) e-mini futures contracts, triggering cascading program-trading that resulted in a 9% drop of the overall market. Essentially, the initial sale triggered various computer models to hit critical levels. When these levels were hit, sales were entered, driving prices lower. The mythical self-fulfilling prophecy came to Wall Street and drove things crazy long enough to wreak havoc.
There was no intentional wrongdoing detected in the first flash crash, but the impact was felt by many. Trades in certain names were canceled and deemed "erroneous." The problem was that many traders bought and sold as prices spiked, and not all trades were canceled. Depending on which trades were allowed to stand for a given name, you might have made a lot or lost a lot, depending on the level selected for trade cancellation. After the event, new circuit-breakers were put in place to halt trading if too rapid a decline occurred.
How was this time different?
The second flash crash was caused by the intentional manipulation of information when hackers used the Associated Press' Twitter account to post a fictitious report about two explosions at the White House that injured President Obama. The S&P 500 fell nearly 1% in three minutes, even though it ultimately closed higher by more than 1%. It is too soon to determine what action, if any, regulators will take either to adjust circuit breakers or to tighten the flow of information from the major social-media sites, but these are the developments that could have the most lasting impact.
The reaction to the bogus report demonstrates that not only the flash crash, but memories of September 11, have had a lasting impact. The reaction to the terrorist attacks on September 11 was slow, with only a limited number of traders reacting before trading was halted. Those who immediately sold on the news were rewarded for their vigilance, but a dangerous precedent was set: Terrorism news is likely to incite panic-selling before facts are checked.
Perhaps the most interesting, and arguably tragic, observation that can be made is that where news of an attack on the White House caused an immediate reaction, the market reacted rather tamely to the Boston attacks. While the impact of a White House attack would be drastically greater, it is worth considering whether any structural changes are suggested based on the Twitter-crash.
After the all-clear
In the aftermath of another rollercoaster dip in the market caused by an aberration, the question is what regulators should do. Tightening circuit-breakers might be an initial idea, particularly among nonprofessional traders; the amount of wealth that vanished in three minutes seems an unfathomable toll for a prank. The problem with this reaction is that at some point, if controls are too tight, the price-discovery function of the market is lost. The free market only works if it is free. A 1% negative spike is significant, but it should not cause the same amount of concern as the 9% dip that resulted from the first flash crash. Ultimately, leaving things alone makes more sense.
Perhaps the bigger concern is of how information is handled. While there is little chance that regulators will be able to prevent the flow of news through social media, there is the possibility that they will address automated trading based on these feeds. An anecdotal cause of flash crash 2.0 was selling caused by programs that are triggered by algorithms that read news and react accordingly. For example, if the correct combination of "terrorist," "explosion," and "White House" are detected enough times in certain feeds, sell orders are placed. There is a real possibility that regulators may adjust the rules surrounding these types of automated trading, making the argument that the systematic risk created outweighs the freedom typically given to market participants. This is a slippery slope, and regulators are prone to overreaching.
Overall, seeing how the investigation plays out will reveal the true significance of flash crash 2.0 and where we might see things change as a result.