Biologists create ‘universal algorithm’ to forecast disaster that could help stock investors cash in ahead of a market crisis

A research team of biologists from China and Japan use experiments with mice to show how impending collapses can be predicted in unrelated fields ranging from financial markets to climatic systems using a one-size-fits-all model

PUBLISHED : Tuesday, 29 December, 2015, 3:51pm
UPDATED : Tuesday, 29 December, 2015, 4:38pm

Biologists are among the last group of people that investors head to for financial advice, but a new study from China suggests they may be able to predict where a market is heading more accurately than many wealth managers.

Specifically, this hinges on a universal algorithm created by a team of research biologists from China and Japan which they claim can detect tipping points and predict disasters ranging from financial market crashes to possible regime change.

Their findings were based on experiments conducted on lab rats exposed to toxic gases.

In their paper published recently in the journal Scientific Reports , the team demonstrated how the financial mayhem that stemmed from the bankruptcy of Lehman Brothers in 2008, and other such momentous changes, could be forecast in the same way and with as much accuracy as the time of death of a mouse exposed to a poisonous gas.

The algorithm detected mathematically identical signals hours before the death of the mouse and before the collapse of the US financial services firm.

It was intended to break the barrier between financial analysis and biological science by developing a one-size-fits-all way of forecasting negative events in multiple fields, the team said.

It could potentially “detect regime shifts or the collapse of an economic, climatic or biological system, or the collapse of a financial market,” the authors wrote.

It can also be applied to any pre-existing way of forecasting events of magnitude to make the results more accurate, the team said.

Most complex systems hit a point known as “critical slowing down”. This could be a noticeable drop in an animal’s blood oxygen level just before it dies, or the sudden decline of a company’s stock price just before the firm declares bankruptcy.

However, these momentous events are usually hard to detect because of various forms of “noise” or interference, for example new government policies or even a civil war.

Many models have been developed to forecast future trends by detecting so-called tipping points, but the latest method represents an important step forward as it can see its way past all the noise - something economists and mathematicians have previously not been able to do.

The research team was led by Professor Chen Luonan at the Shanghai Institute for Biological Sciences and Professor Kazuyuki Aihara at the University of Tokyo.

They followed a similar methodology to the one that Einstein - who introduced time as a fourth dimension - used in the world of physics to throw light on a number of unsolved puzzles. Now physicists keep adding more dimensions to try and explain the secrets of the universe.

Inspired by this and the world of biological science, Chen and his colleagues created an algorithm that can increase the number of “dimensions” of a forecasting model.

They found that the more dimensions a model had, the more effectively it could reduce major forms of interference and detect critical changes that lie ahead.

But Yang Jingping, a professor of financial mathematics at Peking University, said it was unrealistic to compare investors in financial markets to mice in a gas chamber in a controlled experiment.

While the mice were unable to escape, the investors could adjust their portfolios at will, said Yang, who was not involved in the research.

If the new algorithm worked, investors would use it to avoid risk, which would alter the outcome of events in the real world, he added.

For example, if a majority of investors listened to a report issued on a Monday suggesting that a listed company would go bankrupt that Friday, they would likely offload their stock and probably cause the company to go under even sooner, thus making the original forecast inaccurate.