Join our community of smart investors

Yes, animal spirits matter

Experimental evidence shows that share prices are indeed driven by sentiment and not just rationality.
May 22, 2018

Maynard Keynes was right to say that share prices are driven by emotions rather than rationality, according to new research.

He claimed that stock markets “will be subject to waves of optimistic and pessimistic sentiment”, because prices are set by animal spirits rather than by a cool-headed rational assessment of future earnings.

You might think that the many speculative bubbles we’ve seen since he wrote that in 1936 prove him right. They don’t. Yes, it is the case that share prices have in the past moved by much more than can be justified by subsequent dividends, as Nobel Laureate Robert Shiller pointed out back in 1981. But it does not follow from this that prices are driven by sentiment. It’s possible instead that high prices could be due to reasonable people attaching small probabilities to the chance of future good times, as Cardiff University’s David Meenagh and colleagues have shown. If those good times don’t materialise, prices will have been too high in retrospect. This doesn’t, however, mean investors were irrational. You can be wrong without being irrational – as you are every time you lose the toss of a coin.

Take, for example, Amazon. At the peak of the tech bubble in 1999 it was priced at over $100. When it fell below $10 in the subsequent crash that $100 seemed irrationally high. Viewed from today rather than 2001, however, Amazon’s $1,600 price makes than $100 seem a bargain. It wasn’t, perhaps, unreasonable for investors in 1999 to think there might have been other mini-Amazons. Yes, they were wrong. But not necessarily irrational.

Actual share prices, therefore, don’t give us conclusive proof that animal spirits and sentiment drive prices.

But we have other evidence – from laboratory experiments. The great thing about these is that economists can design artificial assets that have known terminal values and known future dividends. This means their fundamental value can be known for sure, which isn’t the case for real world shares. By getting subjects to trade these assets under laboratory conditions, economists can test for mis-pricing.

Tilburg University’s David Schindler and colleagues have done just this. They found that assets tend to be overpriced. This corroborates lots of other research which shows that bubbles are common in the laboratory.

Then, however, they did something different. They got some subjects to take a test requiring them to concentrate hard, with the intention of depleting their self-control. They found that when these people with less self-control traded the asset there was a “significantly higher level of over-pricing”. This, they show, is because those with less self-control are more emotional and impulsive: they reported more joy, fear, surprise and enjoyment during trading.

This is strong evidence that animal spirits matter for prices. When these are not reined in, bubbles are more likely.

Of course, the question to ask of any experiment is: is this true outside the laboratory? Does it have external validity.

We’ve reason to think so. If self-control can be depleted so easily under laboratory conditions, it’s even more likely to be run down by the hassles of everyday life: sleep deprivation, stressful commutes, demanding jobs, colleagues and family yapping away at us and so on. And if the small stakes of laboratory trading can generate emotions how much more likely is this to be the case in real world trading with bigger stakes?

What’s more, Professor Schindler’s experiment removes some channels through which animal spirits might operate, such as the tendency for investors to follow each other, or for recent profits to encourage greater risk-taking and overconfidence, at least among men. Such mechanisms mean that sentiment might matter more in the real world than in controlled laboratory conditions. 

You might think that traders lacking self-control would lose money and so be driven out of the market by more rational and disciplined investors. Not in Professor Schindler’s experiments they weren’t. He shows that they on average made much the same profits as those traders who had better self-control. This is because although they were more likely to buy overpriced assets they sometimes sold these at a profit.

Again, there’s a real world counterpart to this. During the tech bubble of 1999 value investors such as Warren Buffett and the late Tony Dye avoided tech stocks and so suffered losses while those who got carried away made money, at least for a while. As Insead’s Bernard Dumas has shown, there’s nothing much that rational investors can do to exploit or correct excess volatility, except in the very long run.

Competition, therefore, does not drive out irrational traders. Which is evidence in favour of Keynes’ (apocryphal) claim that “The market can remain irrational longer than you can remain solvent”.