ArXiv Blog - A verdade está lá fora
The Physics arXiv Blog produces daily coverage of the best new ideas from an online forum called thePhysics arXiv on which scientists post early versions of their latest ideas. Contact me at KentuckyFC @ arxivblog.comARXIV BLOG
Recently on the arXiv blog...
» Want 'n' waste » Rethinking Energy Use in Data Centers » How to Create Quantum Superpositions of Living Things » Econophysicist Predicts Date of Chinese Stock Market Collapse--Part III » Rogue Microwaves Could Help Explain Freak Ocean Waves » Plasmonic Laser Heralds New Generation of Computing » Vapours 'n' wares » Water Vapor's Counterintuitive Effect on Climate » Astronomers Turn To Omniscopes For Low-Cost Observation » First Hot Ice Computer Created
Econophysicist Predicts Date of Chinese Stock Market Collapse--Part III
Posted: 08 Sep 2009 09:10 PM PDT
First came the prediction, then the prediction came true. Now the econophysicist explains how he did it.
Back in July, econophysicist Didier Sornette at the Swiss Federal Institute of Technology in Zurich and a few palspredicted the impending collapse of the Shanghai Composite stock market index.
Their evidence was that the index had been growing at a faster than exponential rate, which was clearly measurable and obviously unsustainable. But Sornette and co were less forthcoming about how they were able to say a collapse was imminent. I was sceptical of this prediction and said so.
And yet at the beginning of August, the Shanghai Composite dropped by about 20 per cent and I was forced to eat my words, puzzled though I still was to their method. .
Today, Sornette and co publish the method they use to make their prediction. Their method is a synthesis of ideas from the economic theory of rational expectation bubbles, from the imitation and herding behaviour of investors and traders and from the mathematical and statistical physics of bifurcations and phase transitions.
From this, they say they have discovered an unambiguous signature of a bubble market about to collapse in the form of super-exponential growth decorated with logarithmic oscillations as in the diagram above. When the oscillations become
Sornette says he has used this method to predict the collapse of several bubbles in the last few years, including the collapse of the oil bubble last year and the US housing market in 2006.
That's important work which could have a profound impact on economic models. If Sornette and his colleagues have put their money where there mouths are, they ought to be super rich by now.
One interesting aspect of Sornette's predictions is that the team claim that they will not affect the market in a way that changes the forecast.
"Even in the presence of investors fully informed of the presence of the bubble and with the knowledge of its end date, it remains rational to stay invested in the market to garner very large returns since the risk of a crash remains finite."
Of course, the end of a bubble doesn't always imply a crash, just a change from super exponential growth to some other kind of market dynamics.
But if Sornette's forecasts prove reliable, the herding behaviour of traders and investors makes it seem inevitable that they will trigger crashes. If that happens, they will become self-fulfilling prophecies that will generate some interesting dilemmas for traders and regulators alike.
Ref: arxiv.org/abs/0909.1007: Bubble Diagnosis and Prediction of the 2005-2007 and 2008-2009 Chinese stock market bubbles
Comentários