Join our community of smart investors
OPINION

Creep

Creep
February 7, 2014
Creep

Didier Sornette and Peter Cauwels at the Swiss Federal Institute of Technology liken markets to the idea of creep in materials science. This says that if you place some metals or concrete under some constant level of stress the material might for a long time appear stable and robust - until it suddenly breaks. "Hardly anything in life is really stable and in equilibrium, but rather slowly creeps towards a change," they say. However, we fail to appreciate this because we think of economics as linear processes - small, gradual changes - when in fact they are non-linear.

In fact, lots of crashes in financial markets are creep-like, with stability suddenly and unexpectedly giving way to crises, for example:

■ For months, investors thought tapering wouldn't much hurt emerging markets - until it suddenly did.

■ In the 2000s, banks' high gearing and reliance upon wholesale funding didn't much worry investors - until it suddenly did.

■ High government debt in southern Europe seemed stable and sustainable in the 2000s - until it suddenly wasn't.

Financial markets, then, are brittle; they might seem stable - until they break. (In fact, organisations can be brittle too - but that's another story.)

But what's the mechanism that causes this shift from stability to crisis? The obvious candidate lies in a form of coordination game. Each individual trader is trying to anticipate the actions of others; he wants to sell before others sell, and buy before others buy. Usually, the outcome of this game will be rough stability, as some sell and some buy. But, occasionally, traders coordinate upon selling, and so markets crash.

In this sense, sell-offs are just like riots. Riots happen when potential trouble-makers correctly anticipate that others will become violent, in which case they feel they can get away with looting and attacking the police. Just as traders sometimes all sell, so people on the streets sometimes all become disorderly. Sure, there might be a trigger for the shift to selling or rioting, but this can usually only be seen with hindsight. Riots and crashes are both unpredictable shifts from order and stability to disorder. They are both creep-like processes.

However, here the analogy between markets and metal fatigue breaks down, in two ways.

First, the properties of metals under stress are in principle knowable. But the properties of social systems, of which markets are a subset, are not because they depend upon the interactions between agents which can change from second to second; society isn't just complicated, but complex.

Secondly, the civil engineer who predicts that a building will collapse sometime is performing a useful service, whereas the economist who warns of a financial crisis sometime isn't: he's just starting the obvious. In markets, what matters is not being right, but being right at the right time. The man who forecast a banking collapse or European debt crisis in, say, 2003, would have lost his clients a lot of money.

Herein, though, lies something curious. You might think that this perspective - with its stress upon complexity, disorder and non-linear processes - contradicts orthodox economics which stresses linearity and equilibrium.

Maybe so, if you're interested in bigthink. But in fact, both perspectives have the same implication for investors - that crashes are inherently unpredictable. As Capital Economics' Andrew Lilico has pointed out, conventional efficient market theory says crashes are unforecastable because if investors did anticipate them, prices would fall beforehand, which means that sharp price falls can only be the result of surprising news. Our alternative perspective, though, also says crashes are unpredictable because they arise from complex non-linear processes.

The message for investors is the same. You cannot preserve your wealth by futurology - a fact to which we are blinded by the vested interests of 'experts' and wishful thinking - but only by good diversification.