Twitter Tuesday shows US equity market resilience

Tuesday's Twitter-inspired mini-crash confirmed the US equity market's vulnerability to misinformation in the social media age, but a rapid rebound suggested the resilience of trading systems, industry experts have stated.

Tuesday’s Twitter-inspired mini-crash confirmed the US equity market’s vulnerability to misinformation in the social media age, but a rapid rebound suggested the resilience of trading systems, industry experts have stated.

Shortly after 1pm, a tweet from the hacked Twitter account of US newswire Associated Press, which claimed an explosion at the White House had injured President Obama, sent US markets tumbling. The Dow Jones Industrial Average slipped 145 points, or 1%, but the market rebounded to pre-tweet levels four minutes later. 

Matt Samelson, principal at US markets consultancy Woodbine Associates, said the market’s swift reaction highlighted both the problems of using social media to inform automated trading systems, but also the speed of algos in reacting to news that the tweet was false.

“In this age of social media, information is released very quickly before it can be verified as gossip or fact, and markets react just as quickly to misinformation as real fact,” Samelson told today.

Circuit breakers, implemented in the wake of the 6 May 2010 flash crash, which exist to contain widespread panic in markets, were not set off because their limits – around 10% – were far from met.

“There’s no ‘right’ limit a circuit breaker should have – their core function is to avoid widespread panic in markets, and you need some latitude within these systems because your want markets to remain active,” said Samelson.

Social media information is used in some forms of algorithmic trading through automated news reading tools, which scan Twitter, blogs and news articles to ascertain the impact of breaking news on stock prices. In some cases, information is put through sentiment engines, which determine whether news has a positive, neutral or negative effect, either on the market as a whole or on specific securities. Such tools then send the appropriate buy or sell signal to an execution algorithm.

Joe Gits, president of Social Market Analytics, a firm set up to feed social media information into algorithms, said yesterday’s Twitter crash showed the challenges of re-tuning markets in the digital age.

Gits said yesterday’s blip was a hacking, not a social media issue.

“It’s a challenge for automated news reading tools to filter a tweet in this situation because it was from a legitimate Twitter account. If it had come from an individual’s account it wouldn’t have had an impact, but the market reacted because AP is an acknowledged news source, even in tweet form,” Gits said.

“Tweets usually go through a two-stage filter, checking firstly the account, then a sentiment engine, and this would have scored very badly, which has led to algos selling positions and the market going down so quickly,” Gits, a co-founder of database solutions provider Quantitative Analytics, said.

Although Gits said some automated trading programmes may have been able to generate a profit from the small-scale crash by automatically selling in response to the fall in values and buying in response to the uptick.