Simple correlation that links blogs and tweets to Dow Jones average


  • English
  • Arabic

A new zeitgeist is doing the rounds in academic circles these days, and if we did not see it coming, we are obviously not meeting the right people.

Some call it "socionomics", others "mood mining". In any event, it's about the hottest subject around - social media and their profound effect on world markets.

In academic terms: is the public mood correlated with or even predictive of economic indicators?

It seems that those blogs, tweets and Facebook pages are not just random cyber chitchat, but are edged with gold. They can be used to predict market sentiment and make money. It works like this - a surfeit of happy tweets means the market is likely to go up and a glut of unhappy tweets means there is likely to be a bloodbath.

The geniuses that found the correlation between the Twitterverse and the Dow Jones Industrial Average published a paper late last year that explained how they did it. The problem is, while they have produced an amazingly incomprehensible algorithm that claims an 87 per cent correlation, they cannot explain how it works.

The paper, entitled Twitter Mood Predicts the Stockmarket, is the product of a think tank at Indiana University.

Two of the authors, Huina Mao and Xiao-Jun Zeng, are high-ranking engineers and mathematicians, and the third, Professor Johan Bollen, is a social scientist cum psychologist. They are not experienced market traders or investors, but if they were, they might have saved themselves the trouble and talked to their local stockbroker.

But that would be too easy. The original questionnaire, which was developed in 2008, asked people to rate how closely their feelings matched 72 different adjectives, including "friendly", "peeved", "active", "on edge" and "panicky", and uses the responses to measure mood along six dimensions: calmness, alertness, sureness, vitality, kindness and happiness. "We're using Twitter like a psychiatric patient," Prof Bollen says. "This allows us to measure the mood of the public over these six different mood states."

No prizes for guessing that when the mood on the Twitterverse was buoyant, markets surged. They even checked it out with empirical accuracy by correlating the market on known days of happiness. Around the time of Barack Obama's inauguration as US president and on Thanksgiving Day 2008, the markets soared.

Anyone with half an eye on global equities over the past two decades will have heard something like this before.

Everything from the weather to the lunar cycles have been used to predict human emotions, and by extension, stockmarket sentiment. In the 1990s, a London fund management group called Gaiacorp (Gaia equals Greek goddess of the Earth) managed to convince its clients that currency gyrations followed the flow of the tides and backed it up with an unbeatable algorithm as well.

Still, there are the believers - Derwent Capital Markets, a London hedge-fund outfit, has placed $US40 million (Dh146.9m) of clients' money for exclusive rights to Prof Bollen's Twitter predictor algorithm. To his credit, he says he is not the first to have a technique like this. Blogs and news have been "mood indicators" for some time, he notes. He also says he wants to put "his money where his mouth is" and do this "in real time".

Whether you short the market when the public sours, or go long when happiness abounds, you are trading people's emotions, which is tantamount to trading on pure chaos. There are some who do this already, but they do not purport to be able to predict the public mood, only to trade on its swings and roundabouts.

Anything that can be measured can be traded - and volatility, as an entity, is measurable. The Chicago Board Options Exchange Market's volatility index (known as the VIX) works on measuring the implied volatility of S&P 500 call options, and is often termed the "fear and greed index".

People can now take a view on volatility itself - and exchanges around the globe are now giving thought to publishing indexes that reflect the volatility of shares.

"Wherever there's an index, there's a futures contract over that index," a trader tells me. "We can trade the underlying futures contract and we can make money when things get more volatile."

Chaos, it seems, is just another asset class.