Robert Shiller says information cascades can explain asset bubbles:
How a Bubble Stayed Under the Radar, by Robert Shiller, Economic View, NY Times: One great puzzle about the recent housing bubble is why even most experts didn’t recognize the bubble as it was forming. ...
Were all these people stupid? It can’t be. We have to consider the possibility that perfectly rational people can get caught up in a bubble. In this connection, it is helpful to refer to an important bit of economic theory about herd behavior.
Three economists, Sushil Bikhchandani, David Hirshleifer and Ivo Welch, in a classic 1992 article, defined what they call “information cascades” that can lead people into serious error. ...
Mr. Bikhchandani and his co-authors present this example: Suppose that a group of individuals must make an important decision, based on useful but incomplete information. Each one of them has ... information..., but the information is incomplete and “noisy” and does not always point to the right conclusion.
Let’s update the example...: The individuals in the group must each decide whether real estate is a terrific investment... Suppose that there is a 60 percent probability that any one person’s information will lead to the right decision. ...
Each person makes decisions individually, sequentially, and reveals ... decisions through actions — in this case, by entering the housing market and bidding up home prices.
Suppose houses are really of low investment value, but the first person to make a decision reaches the wrong conclusion (which happens, as we have assumed, 40 percent of the time). The first person, A, pays a high price for a home, thus signaling to others that houses are a good investment.
The second person, B, has no problem if his own data seem to confirm the information provided by A’s willingness to pay a high price. But B faces a quandary if his own information seems to contradict A’s judgment. In that case, B would conclude that he has no worthwhile information, and so he must make an arbitrary decision — say, by flipping a coin to decide whether to buy a house.
The result is that even if houses are of low investment value, we may now have two people who make purchasing decisions that reveal their conclusion that houses are a good investment.
As others make purchases at rising prices, more and more people will conclude that these buyers’ information about the market outweighs their own.
Mr. Bikhchandani and his co-authors worked out this rational herding story carefully, and their results show that the probability of the cascade leading to an incorrect assumption is 37 percent. ... Thus, we should expect to see cascades driving our thinking from time to time, even when everyone is absolutely rational and calculating.
This theory poses a major challenge to the “efficient markets” view of the world... The efficient-markets view holds that the market is wiser than any individual: in aggregate, the market will come to the correct decision. But the theory is flawed because it does not recognize that people must rely on the judgments of others. ...
It is clear that just such an information cascade helped to create the housing bubble. And it is now possible that a downward cascade will develop — in which rational individuals become excessively pessimistic as they see others bidding down home prices to abnormally low levels.
Here's another example of how this mechanism can work: Suppose you are in a long line of cars that are all headed to the same place. You have a vague, but not perfect idea of how to get to your destination. At one intersection, by chance, the first three cars get it wrong and turn left instead of right. The fourth car, though somewhat certain the correct way to go is to turn right, is not certain enough to go in a different direction and chooses to follow the first three cars. The driver assumes that since all three cars turned left, that must be the right way to go. As everyone begins to turn left, the signal strengthens and only those who are very certain of the true direction (which is a difficult condition to meet in asset markets) will choose to deviate from the group and go in the direction they know is correct. [For this to work, it is important that decisions be sequential, and that people can observe and learn from the decisions of others.]
We don't know for sure how or why bubbles occur, so this may or may not be the correct explanation. In general, it's hard for economists to admit they can occur at all since they do not appear to be rational responses to economic fundamentals and we do not have very good models of them, so this may explain some of the resistance to admitting a bubble is occurring at the time it is inflating. We prefer to explain market outcomes with economic models that we understand instead of with bubbles that occur for mysterious reasons. It is only after the fact, when it becomes clear that no fundamentals based interpretation seems possible, that we admit a bubble could be the correct explanation. But maybe recent experience will change that and we won't be so quick to deny it when the next bubble begins to inflate, though I can't say I'm very confident in that prediction.