Back in 1936 Maynard Keynes said that share prices were more volatile than they should be. “Day-to-day fluctuations in the profits of existing investments, which are obviously of an ephemeral and non-significant character, tend to have an altogether excessive, and even an absurd, influence on the market,” he wrote. Prices, he continued, are “liable to change violently as the result of a sudden fluctuation of opinion.”
This, though, was no more than an idle claim. He offered no evidence to support it, other than a passing reference to American ice manufacturers and UK railway shares. It was a mere hypothesis.
Findings that earned a Nobel Prize
Then in 1981 Robert Shiller – in work which was to win a Nobel prize – provided some solid support for this. He calculated that between 1871 and 1971 the S&P 500 was between five and 13 times more volatile than could be justified by changes in expected dividends.
This work represented genuine progress. An uncorroborated claim became an accepted empirical finding.
But our story doesn’t end there. Why are shares more volatile than dividends? The inference which Shiller invited was that investors were simply irrational. But he didn’t provide direct evidence for this. And this invited another interpretation of his result. If investors attach reasonable and varying probabilities to events which could have happened but did not, then prices will seem excessively volatile in hindsight even though they are reasonable at the time. For example, prices now seem to have been “too low” in the 1930s and 70s. But was this because investors were irrationally pessimistic then, or was it because they thought – reasonably but with hindsight wrongly – that there was a chance that capitalism would collapse and so avoided shares?
Do investors imitate each other?
To adjudicate here, we need to study individual investors’ behaviour. And recently, economists have been doing just this. Matteo Ploner at the University of Trento ran some experiments in which subjects were asked to choose between safe and risky investments. He found that when subjects were told what others were doing, their choices came to resemble others. "Social effects play a fundamental role in investment choices" he concluded.
This suggests an explanation for Shiller’s result. If investors imitate each other, then some will buy when prices are high because others are doing so. In doing so, they’ll drive prices even higher.
One problem with experimental evidence in economics is that it raises the issue of external validity: is what’s true in the laboratory true of the real world? In this case, we have a reason to think so. Hans Hvide at the University of Aberdeen and Per Ostberg of the University of Zurich have found that people are disproportionately likely to buy shares if their colleagues do so. Peer effects, then, exist in the real world too.
And they are not necessarily irrational. Imagine you’re looking for something to eat in a strange town. You find two restaurants: one quite full, the other empty. Which do you choose? It would be entirely reasonable to go to the fuller one. The locals, you figure, could know something you don’t. Similarly, when faced with uncertainty about the investment outlook, why not do what others are doing?
A reason behind bubbles and slumps
What we have here, then, is a story of progress. We’ve gone from a vague claim that share prices are too volatile, through good empirical evidence that they are, through to a well-founded theory of why they might be, based upon studies of individual investors.
In this, we’ve learned two things. One is that the imitative behaviour that causes shares to become too cheap or too expensive needn’t be outright irrational. It might instead arise from people trying to learn in an uncertain environment.
The other is that we shouldn’t think of share prices as if they were the opinion of any single individual. Instead, they are the outcome of inter-actions between people. If investors imitate each other – so that buying leads others to buy – we can get bubbles and slumps. And if they don’t, we’ll get more stable prices.
Sadly, though, we cannot yet forecast when imitation will be strong and when it won’t. The progress of knowledge has not, therefore, given us the ability to make profitable predictions. But then, the idea that it must do so owes more to ideology and wishful thinking than it does to the evidence.