Financial Flocking

Wonderfully Herd-like feature in this weeks New Scientist talking about the inner workings of financial markets and theorising about the inadequacies of prevalent economic theory and the potential real causes behind the current financial malaise.

Traditional economic theory posits that markets operate in 'equilibrium'. They are affected by changes in surrounding forces – good news about a company creates increased demand for its stock and the price goes up, bad news decreases demand and it goes down. Financial markets are shaped by the tendency of company stocks to find their 'proper' values, driven by the incentive of investors to identify the real value so that they don't pay too much for the stock or sell it for too little. With millions of investors acting in this way, any stock soon finds its 'true' value and any mispricing (however temporary) is corrected by market forces. In this way, drammatic swings in the market can only follow from correspondingly drammatic causes such as a significant piece of news, good or bad.

Sounds logical, right? Except that its not. Physicist Jean-Philippe Bouchand analysed over 90,000 news items over two years (relevant to hundreds of stocks) from Reuters and Dow Jones who produce real-time news feeds to investors. They were looking for the direct link between the news stories and corresponding jumps in stock price. There wasn't one. They found that "neither idiosyncratic news nor market wide news can explain the frequency and amplitude of price jumps". Instead jumps seemed to "occur for no identifiable reason". The markets seemed to have an internal dynamic all their own.

This is born out by recent experience. The 2007 global plunge in stocks seemed to be characterised by the emotive forces which drove it, with words like 'fear' being often used as a situational descriptive at the time. When, in a previous study, Bouchand looked at the accuracy of analysts market predictions he found that not only did they have a tendency to be over-optimistic but that their predictions were often similar to those already made public by other analysts even when this went against prevailing information.

And contrary to the normal distribution of random events characterised by the bell curve financial fluctuations have a tendency to have 'fat tails', meaning that large price fluctuations are more likely than expected to occur and their likelihood often underestimated by those employing equilibrium thinking. The New Scientist argues that perhaps we should have seen this coming:

"Some ecomonists have long argued that the movement of opinions and information between people tends to amplify market movements, leading inevitably to fat tails."

Paradoxically, it can be the attempt of investors to learn the relevance of new information (often by watching others) that can amplify price swings. Work done by Didier Sornette, an econophysicist at the Swiss Federal Institute of Technology, showed that whilst public and private information tended to keep prices around realistic values, the information which flows through social networks and gets spread by word of mouth tends to create groups of people co-ordinated in their actions "which in turn leads to bubbles – stocks that become priced too high or low". These bubbles can be triggered by random pieces of news which get amplified through the social network. In this way, rather than being precedented by a major event, market crashes can have more to do with "a progressive linking together of investors' decisions and expectations over months or years". This contributes to market instability, leading to a high likelihood of a crash occuring, which can eventually be triggered by relatively minor contributory factors.

Image courtesy

A new generation of financial market simulations is attempting to take account of this 'financial flocking' – including detail on what makes people trade and how the actions of one investor can influence others, and even going so far as building computer models to predict market dynamics populated by artificially intelligent 'agents' that mimick the activity of real markets. Such models have had some impressive successes at reproducing stock histories.

But it is notable that as highly leveraged trades (notably hedge funds borrowing from the banks) increase the amount of leverage in the market, the risk of a cascade of failures increases exponentially with higher levels of credit creating stronger links between players in the market which in turn heightens the avalanche effect.

So it is perhaps the increasing access to easy money which is the single most destabilising factor. As the article points out, it is common for financial incentives to induce people to act for their own short-term benefit whilst saddling someone else (the client, or the company they work for) with the longer-term risks. As the sub-prime situation has shown, this thinking works fine if you believe in equilibrium economics. It's entirely another matter however, if you believe in the power of the rapid spread of ideas amongst groups of people.

Full article here