Sharing knowledge and information can help to predict accurately the fluctuations in financial markets.

Traders forecast market movements much better when they discuss together than when they try to predict in isolation. This positive consequence of pooling information has been highlighted in a study carried out by a team of three researchers from the United States and Spain*. They looked at an instant messaging network installed in a financial sector company where traders worked. The researchers were seeking to explain why financial traders who communicated with each other obtained far more positive financial results than those who chose to simply rely on their own judgement.

Collective knowledge - a synthesis of multi-individual knowledge

For their study, the researchers used a traditional method, i.e. observing how often words which are normally not often used were in fact used on this discussion platform. Once they had identified these key words, the researchers went on to identify the discussion topics in which they appeared and their frequency of use, and then examined the corresponding market fluctuations. Having analysed a number of discussions, the researchers concluded that a kind of "collective knowledge" had emerged from the inter-changes. It seems that whenever different opinions – whether right or wrong – come up against one another, this always leads the traders to a better understanding of the markets than when they just think things through by themselves.

Whether short- or long-term, the predictions prove highly effective

Basically the phenomenon which the researchers observed amounts to a kind of disappearance of differences of opinion. The differences fade away, merging into an overall opinion which is often very close to the true picture. On days when the market is undergoing rapid, violent fluctuations, the trend which emerges is a fusion of the most talked about topics. On the other hand, a somewhat calmer market will encourage traders to try to predict movements during the days ahead. The researchers point out that, given that the key words used are not predefined, this process can be replicated across all business sectors.

*Serguei Saavedra, Jordi Duch, and Brian Uzzi, from Northwestern University and Rovira i Virgili University