Debt, stocks, restaurant, economics, portfolio, religion, cancer. You can sort of see the link between some of those words, but not all. But there is a link apparently, and the link is with stock market performance.

Take the word debt. If there is a decrease in search engine traffic using this word, it would appear it is a good time to buy into stocks. If there is an increase in the amount of times users type the word debt into Google, it is a good time to sell – or so suggests a new report published in ‘Nature’s’ scientific reports, by Tobias Preis, (Professor at Warwick Business School) and Helen Susannah Moa and H. Eugene Stanley from Boston University.

Now take the word ring. It appears the relationship is the other way round. If search traffic rises, then it may be better to bail into stocks. Of course this is not proven, and there are reasons to question the findings, but let’s run with the story for a bit longer.

The three academics looked at 98 search terms. They first concluded that rises and falls in search engine traffic of certain words were correlated with stock market performance. This is no big deal. After all, knowing that more people type the word debt into Google when stocks are falling doesn’t help that much. What investors want is to be able to predict changes in the stock market, not have a new way of describing them.

But the academics also found that certain key words did rise and fall in popularity before changes in stock prices. They found the model worked better when applied to local stock markets. So, for example, tracking search terms used in the US as a way of predicting changes in US stocks was more reliable than using global search traffic to predict global stocks.

This is not the first example of a study appearing to correlate Internet traffic with stock market performance. For example MIT academics Sandy Pentland and Yaniv Altshuler found that investors who are plugged into a diverse range of investment groups enjoy better returns. Last year Johan Bollen and Huina Mao of Indiana University and Xiao-Jun Zeng of the University of Manchester found that investors who tap into the public mood often enjoy superior performances. They also found that Twitter is a good gauge of such mood. See: A tip for investors: embrace social media – but know when to bail-out


So does the study stack up? Does it make sense?

There has been no shortage of critics. But, from one point of view it does surely make sense. Crowds are complicated things. Psychologists have shown we all tend to comply with the crowd, but – and this is a subtle point – maybe it goes further than that.

Different people often tend to react in the same way in the same circumstances. Who knows why certain type of clothes or child names go in and out of fashion, but they do.
The mood of the crowd can surely relate to stock prices. It can show when the crowd is in the mood for buying, and can surely predict when it is getting nervous.

Can such models warn of one of the black swan events – that is to say rare, but highly significant events? Well, the finance crisis of 2008 was a black swan event if ever there was one, and yet the study by Preis, Moa and Stanley covered data from 2004 to 2011. In short, the occurrence of black swan event did not validate their findings.

And yet, sometimes the mood of the crowd can create problems it is unaware of until the moment the problem becomes irreversible. That is the essence of the book ‘The Blindfolded Masochist’ by yours truly. Intuitively, this aspect of crowd behaviour should mean that on occasions Google will fail to predict changes in stock prices.

The other drawback is that such systems can break down when they are commonly used. So if all big investors incorporated Google analysis into their investment strategy, then such a strategy would surely back-fire.

For the report in full, see: Quantifying Trading Behavior in Financial Markets Using Google Trends 

© Investment & Business News 2013