Yale University's Shyam Sunder and colleagues created an artificial market on a computer in which half of traders were well-informed while half were programmed to be ignorant and updated their beliefs about future prices using a simple rule of thumb. They discovered that in this market in which many traders had zero intelligence prices quickly converged to what they would have been if all traders were intelligent.
This, they say, shows that rational markets don't require all traders to be sophisticated. Instead, efficient markets arise from market structure rather than from the characteristics of particular traders.
One complaint about experimental markets such as these is that they might lack external validity: what's true in the laboratory might not be true in the real world. However, this objection might not apply in this case.
For one thing, it's consistent with research by Gerd Gigerenzer at the Max Planck Institute. He's shown that we don't need sophisticated optimisation strategies to make good decisions because simple rules of thumb work as well.
What's more, studies of some very different real-life markets are also consistent with Professor Sunder's finding. Alan Kirman at the University of Aix-Marseilles looked at fish markets in Marseiiles and Ancona. He found that for many individual buyers, demand curves didn't slope downwards; they didn't buy more fish when prices were lower. For the whole markets, however, they did.
This means that markets are emergent processes, in the sense that their outcomes are the unintended result of interactions between individuals, and not individual behaviour write large. "Aggregate behaviour does not always have a counterpart in the microeconomic data," says Professor Kirman.
Although emergence is a relatively new field in economics, this view is in fact an old one. When Adam Smith wrote in 1776 that "it is not from the benevolence of the butcher, the brewer, or the baker, that we expect our dinner, but from their regard to their own interest" he was describing how selfish individuals can produce benign outcomes. That's emergence.
However, emergence works both ways. Just as stupid people can produce intelligent markets if structures are right, so intelligent ones can produce stupid markets if structures are wrong. One reason why some fear that equity markets are not yet fully pricing in a rise in US interest rates is that otherwise smart traders are incentivised to have short-term horizons. The man who shorts an over-priced market might lose his job if prices rise before he is proved right.
All this might help explain a paradox pointed out by the late Nobel Laureate Paul Samuelson, that stock markets are micro efficient but macro inefficient; it's hard to outperform the market by picking individual stocks, and yet the aggregate market is quite often over or under priced.
One reason for this is that it is often riskier to short the aggregate market than individual shares. In the latter case, you can accompany a short position with long positions in similar but under priced stocks and so minimise market or sector risk by pairs trading. You cannot, however, go short of whole markets without taking on big risk. In this sense, markets can be micro efficient but macro inefficient not because individual traders are micro rational but macro irrational - which might not even be psychologically possible - but because of differences in the structure of markets at the macro and micro levels.
What matters, therefore, is not so much individual psychology as incentives and market structure. The question: are markets efficient? Is very different from the question: are investors rational?
