We leave an indelible electronic footprint whenever we use the web. In recent years, computer scientists have discovered that this presents them with a remarkable opportunity to record and and track the behaviour of humanity in its virtual incarnation.
One important question is to what extent behaviour in cyberspace influences or reflects behaviour in the real world. There are various examples of close correlations. Perhaps the best is the powerful predictive links that researchers have discovered between flu-related web search terms and the actual spread of the disease.
Today, Matthieu Cristelli at the Institute of Complex Systems in Rome, Italy, and a few pals unveil another example. These guys have trawled the search queries submitted to Yahoo! looking for mentions of companies in the NASDAQ-100 stock market index.
What they’ve found is remarkable. They say the volume of search queries related to companies on the NASDAQ-100 are correlated with the volumes of trades in those same companies in the following days.
That’s surprising because Cristelli and co say that most webusers check each stock of interest just once per month. This implies that they are not expert traders. So the effect emerges from the collective but uncoordinated activity of many inexpert users–a kind of wisdom crowds.
Just why this should be the case is a mystery.
However, the team says its discovery could be extremely useful. “We think that this information can be effectively used to detect early signs of ﬁnancial distress,” say Cristelli and co.
That’s a hypothesis that will need careful testing. There have been many examples of wisdom-of-crowd effects melting into the background noise.
This one needs to be treated with care because there is no mechanism for cause and effect. That’s in contrast to the example given above of search terms related to flu being a predictor of the spread of the actual disease. That’s believable because it’s entirely reasonable to think that people suffering these symptoms might look them up on the web.
But the idea that websurfers with little trading expertise can somehow be responsible for driving the volume of trades in any particular stock is not so easy to accept. Cristelli and co will need to work harder to explain how this could happen. Perhaps the volume of searches reflect some other influence, such as stories in the mainstream news or advertising campaigns, which influence professional and casual traders alike.
Stock picking is a notoriously slippery practice, littered with pitfalls and snake oil. Without some reasonable explanation of cause and effect, critics will find it hard to dismiss the possibility that this result is little more than a crazy correlation thrown up by chance.
Ref: arxiv.org/abs/1110.4784: Web Search Queries Can Predict Stock Market Volumes