Flocking Behaviour Improves Performance Of Financial Traders

Day traders tend to coordinate their behaviour in the same way that cicadas synchronise their chirping

Emerging Technology from the arXivarchive page

October 5, 2011

Cicadas chirp to attract mates but the noise can also bring the unwanted attention of predators. That leads to an interesting synchronised behaviour in which the creatures chirp in a coordinated fashion, creating a kind of chorus.

It’s not so hard to see why. Cicadas that chirp ahead of the chorus stand out. That makes them more likely to attract a mate but also more likely to be eaten by a predator. However, it turns out there is a best time to chirp, a kind of “sweet spot” in time, that maximises the likelihood of reproductive success while minimising the risk of death. That’s what generates the chorus.

Humans also demonstrate synchronised behaviour. One of the most interesting examples is in science when a number of individuals often make the same discovery at the same time. The reason is that those who announce their discoveries too early are likely to be condemned as cranks by their peers, while those that are too late have obviously missed the boat.

Instead, there is a kind of “sweet spot” in time when new discoveries are most likely to be confirmed and welcomed by peers.

Today, Serguei Saavedra and pals at Northwestern University in Evanston, Illinois, reveal another example, this time in the behaviour of day traders who buy and sell stocks.

Saavedra and co analysed the trades of 66 day traders at an anonymous trading firm between 26 September 2007 and 20 February 2009. During this time, these guys made in excess of 1 million trades, making a profit in just over half of them (55 per cent).

On the face of it, the traders have little reason to synchronise their behaviour since they buy and sell different stocks in different sectors, such as healthcare, high tech and so on.

But when the Northwestern team analysed the trades on a second by second level, they found that there were far more synchronised trades than pure chance can account for.

And these trades were more successful too. Traders who successfully synchronise their trades tend to be better off at the end of each day.

The question is how does the synchronisation occur? Saavreda and co say that although each trader works in different areas, the stocks are influenced by certain types of general news such as Federal Reserve announcements, new job ﬁgures,housing market changes, speculation about bankruptcies and so on.

This information is available to all traders at the same time but then has to be digested and its impact evaluated. The same news can have an entirely different impact on different stocks. Traders then have to decide what to trade, how much to trade and, crucially, when to trade.

Out of all this seemingly unconnectedness, synchrony still emerges between entirely unrelated trades. Saavreda and co say that 96 per cent of simultaneous trades–those that occur within a second of each other–are of different stocks.

That points to the existence of a kind of sweet spot for trades, a moment in time when it is best to cut a deal. Traders that hit this sweet spot are more likely to come out on top.

And that’s exactly what Saavreda and co discovered. “We found that when a stock trader in that ﬁrm trades at the same time as other traders in the ﬁrm, his or her ﬁnancial performance is signiﬁcantly increased,” they say.

But here’s an interesting twist. During the working day, traders constantly message each other through a network that, by law, has to be carefully monitored. Traders use this system to make sense of the news they are receiving.

During the period in question, the traders sent more than twice as many messages as they made trades: more than 2 million of them. To look for links between these messages and the trades, Saavreda and co trawled through these too.

It turns out that messaging patterns are positively correlated with the level of synchrony. In other words, as traders exchange more messages, they become more synchronised.

That means the behaviour of traders and cicadas is strangely analogous. Just as chirping is designed to improve reproductive success, messaging can improve trading success. But both work best when they are executed at specific sweet spot in time .

But get this: Saavreda and co found no evidence of any centralised control. So the synchrony is an emergent phenomenon in traders, just as it is with cicadas.

All that suggests some interesting lines of future research. Saavreda and co make the rather unlikely suggestion that looking for similar patterns of synchrony may help national security specialists prevent acts of terrorism or help disease control experts prevent pandemics.

A more interesting approach might be to attempt to recreate this effect in a simulated trading lab and then study the frightening scenario in which the synchrony becomes dysfunctional.

In that way, these guys may get a handle on the process that turns the collective wisdom of crowds into the madness of mob rule.

Ref: arxiv.org/abs/1110.0381: Synchronicity, Instant Messaging And Performance Among Financial Traders

Keep Reading

Large language models can do jaw-dropping things. But nobody knows exactly why.

And that's a problem. Figuring it out is one of the biggest scientific puzzles of our time and a crucial step towards controlling more powerful future models.