Econophysicist Predicts Date of Chinese Stock Market Collapse

The boom and bust nature of economics is one of the most puzzling aspects of the modern world. In the last year or so, many people have learned to their cost that when bubbles burst, businesses, jobs, and livelihoods can go with them.

So an obvious question arises: can we spot bubbles when they occur and predict when they are about to burst? One group of theorists say that they can and have used their techniques to make an extraordinary prediction.

First, they say that they've found the telltale signs of a bubble in the growth rate of the Shanghai Composite stock-market index. And second, they say that this bubble will burst between July 17 and 27.

That's a brave move, so let's look at it in more detail. The theorist behind this prediction is Didier Sornette at the Swiss Federal Institute of Technology, in Zurich, who has pioneered the study of market bubbles. Last year, he used his method for spotting bubbles to reveal that oil prices where dangerously inflated.

The telltale sign of a bubble, he says, is a faster than exponential growth rate caused by a positive feedback mechanism that generates this nonlinear growth.

The faster than exponential growth rate is relatively easy to spot. According to the analysis done by Sornette and a few mates, the Shanghai Composite Index certainly seems to have had a faster than exponential growth--a 69 percent rise since October of last year.

Whether an unsustainable positive feedback mechanism is causing this growth isn't so clear. Sornette and co suggest that what is responsible is the Chinese government's massive lending spree designed to maintain its economic growth rate at 8 percent a year. China has maintained that kind of growth for some years now.

Let's take at face value the idea that a bubble has formed. What of the prediction that it is about to burst? Just how this team arrives at such a precise date isn't clear, but whatever the mechanism, this is a much more speculative move.

One thing that physicists have learned about complex systems, such as stock markets, earthquakes, and forest fires, to name just a few, is that when changes occur they are scale invariant.

That means that if you were to remove the numbers from the axes of a graph plotting this behavior, there is no way that you could identify the scale of the events by looking at the plot. This implies that there is really no difference in principle between a small change in the stock market today and catastrophic change tomorrow.

That makes predictions of almost anything, let alone the imminent collapse of a bubble, extremely hard to make. Impossible may not be too strong a word for it.

Sornette and co do not say how they make their prediction, but they do hedge it by saying, "This will lead to a change in regime which may be a crash or a more gentle bubble deflation."

But they're still predicting an end to this faster than exponential growth in the Shanghai Composite Index between July 17 and 27. That change in growth will of course happen one day, but you'd have to have very good reasons to say that it will occur between those two dates. Those reasons are missing from this paper.

I say that this is a prediction that is impossible to make. And I'm prepared to bet Sornette that he's wrong. The stakes? Let's say an arXiv blog baseball cap and T-shirt.

Ref: arxiv.org/abs/0907.1827: The Chinese Equity Bubble: Ready to Burst