Join our community of smart investors
OPINION

Chasing risk

Chasing risk
August 24, 2016
Chasing risk

Tobias Rotheli at the University of Erfurt got subjects to play a computer game where they chose where to fish on a virtual frozen lake; places furthest from the shore yielded more fish but also increased the risk of the ice breaking causing everyone to lose their fish. The analogy with financial markets is clear: in both cases, there's a choice between a safe and risky option.

And here's the thing. Mr Rotheli found that crashes were more likely when he removed the option of catching fish safely. That removal led people to take more risk and so suffer more crashes - so much so in fact that the risk-return trade-off disappeared because crashes were so common as to offset the good returns when the crashes didn't occur.

This contains a clear warning to investors: when returns on safe strategies are poor, the risk of a crash is greater because people pile into risky assets.

Of course, all experimental research invites the question of external validity: is what happens in the laboratory applicable to the real world? Three things make me believe so in the case.

First, we have abundant evidence from other experiments which confirms that overvaluations and crashes are common. In a survey of experimental asset markets, Charles Noussair at Tilburg University says there is "a strong tendency to generate price bubbles and crashes."

Secondly, other experiments by Harvard University's Carmen Wang corroborate Mr Rotheli's finding. She and her colleagues showed that low interest rates cause a "reach for yield" - the buying of risky assets even though their payoffs haven’t changed. This can increase the chances of a crash as it pushes share prices higher than their fundamentals would justify.

Thirdly, in a world of secular stagnation low interest rates don't stimulate much productive investment simply because there are few profitable projects. Instead, they are more likely to cause a stampede into malinvestments, as we saw in the mid-00s when low yields on safe assets led to over-investment in credit derivatives.

So, does all this mean we can confidently predict a crash?

No. Mr Rotheli's work shows that crashes are not the result of individuals behaving irrationally. Instead, they arise from combinations of behaviour by different people: they are a complex emergent phenomenon. Some combinations lead to price stability - if for example some of those who have enjoyed good returns scale back their risk-taking - while others produce crashes: if one investors chases risk, it can pay off but if many do so it leads to aggregate over-pricing.

The thing about emergent processes is that they are unpredictable. For this reason while we can warn of the possibility of a crash, and even describe the sort of mechanisms that would generate one, we cannot say for sure when one would happen. Some things cannot be known.