Can Search Engines Predict Financial Market Activity ?

By October 27, 2011

Predicting movements in the financial markets several days in advance might be made possible to a certain extent by studying Internet user searches on dedicated sites.

It appears that Internet user searches via engines such as Yahoo! may provide a basis for predicting financial market activity several days ahead. Five European researchers started out from the principle that if it is now possible to forecast to a certain degree the birth of social phenomena by looking at the number of user searches, this might apply to other fields as well. By observing for one year from April 2010 to May 2011 the searches linked to NASDAQ 100 companies on Yahoo! they ascertained that there is indeed a “collective consciousness” on the Web.

Links between Internet searches and financial activity

The researchers noticed that 90% of Internet users only search once per year on a specific company name. This implies that they are not really financial experts but individuals whose investment portfolios contain only one or two types of shares. The researchers also realised that there’s a correlation between the companies which attract a large number of searches at a given moment and the number of shares being traded over the following days. Thus a group of separate non-specialised individuals can facilitate prediction of some market movements.

A way to nip a crisis in the bud…?

This information would lead us to think that, contrary to current opinion, the majority of investor portfolios are not well balanced – since they carry only a very small number of different stocks – which might explain the speed and frequency of domino effects observable in the markets. The researchers are planning to refine their research model, adding data from Twitter and the semantic analysis of blogs. Readers may recall that some months ago a group of researchers from the Universities of Indiana and Manchester managed to determine the Dow Jones Index to 90% accuracy by monitoring the mood of Twitter users as evidenced in their tweets.

Legal mentions © L’Atelier BNP Paribas