One idea in the world of finance is that the volatility of a market is a good measure of the risks it represents. So it’s easy to imagine that volatility should also be a good predictor of financial crises, when the biggest corrections occur.
That’s not the case, say Dion Harman at the New England Complex Systems Institute in Cambridge, MA, and a few buddies. They say that while volatility increases at the beginning of a crisis, it is unreliable as a leading indicator of trouble ahead.
Instead, they’ve found a better predictor of trouble–the presence of sheer, unadulterated panic.
That may sound like a truism but Harman and co say that the tell-tale signs of incipient panic are present well before crises become evident in other ways. And they say they have the evidence to prove it.
First, what is panic? In sociology, panic is defined as the collective flight from a real or imagined threat. So an important element is the way in which individuals copy each other. The critical transition that occurs during a panic is the change from behaviour that is stimulated from outside the group to behaviour that is triggered from within via large scale mimicry.
So one way of measuring the nervousness that precedes panic is to see how closely individuals are copying each other, say Harman and pals. To this end, they measured the fraction of stocks traded on the NYSE or Nasdaq that move in the same direction on particular day.
If these movements are the result of unrelated external stimuli then roughly the same number should move up as down. And sure enough the data from the year 2000 shows exactly this. At any given time in 2000, about half the stocks moved up and the other half moved down.
But throughout the naughties, this fraction changed substantially, say Harman and co. And in 2008, copying behaviour was so ubiquitous that the likelihood of any fraction of the market moving in the same direction was more or less the same. So it was just as likely that 80 per cent of the market would move up on a given day, as it was that 80 per cent would move down or that the split would be 50:50.
Harman and co say this “co-movement” is indicative of a nervous market that is ripe for panic. “The signature we found, the existence of a large probability of co-movement of stocks on any given day, is a measure of systemic risk and vulnerability to self-induced panic,” they say.
And in fact exactly this self -induced panic kicked in when the current financial crisis began in late 2008.
What’s interesting about the approach is that Harman and co say that a similar nervousness can be seen before other big drops in historical stock market data. In fact, when they put their method to work on historial data from the Dow Jones Index, they say their panic-spotting method identifies four year-long periods in which occur 8 of the biggest percentage points drops in the last 26 years.
But it’s all very well spotting the precursors to crises in hindsight. There’s no shortage of people who say they can do that.
The litmus test is whether you can spot them in advance. The list of those able to do that is significantly shorter and arguably an empty set.
For Harman and co to change this they’ll need to take the brave step of making some testable predictions. We’ll be watching.
Ref: arxiv.org/abs/1102.2620: Predicting Economic Market Crises Using Measures Of Collective Panic
We can’t afford to stop solar geoengineering research
It is the wrong time to take this strategy for combating climate change off the table.
A horrifying new AI app swaps women into porn videos with a click
Deepfake researchers have long feared the day this would arrive.
Meet Altos Labs, Silicon Valley’s latest wild bet on living forever
Funders of a deep-pocketed new "rejuvenation" startup are said to include Jeff Bezos and Yuri Milner.
Meta’s new learning algorithm can teach AI to multi-task
The single technique for teaching neural networks multiple skills is a step towards general-purpose AI.
Get the latest updates from
MIT Technology Review
Discover special offers, top stories, upcoming events, and more.