Herding in Financial Markets
In making a decision, individuals have access to private information and public information. We do not have access to others’ private information and so we must use what we can observe as a signal of what their private information might be. When we make a decision that goes against what our own private information, this is likely because we have observed something in the behavior of others that signals they know something important that we do not know. When this happens on a large scale, it becomes an information cascade. To an observer, it can be unclear how much of this effect is based on legitimate information vs. copying the behavior of others.
One particularly relevant example of where this can occur is in financial markets. During both the tech bubble and the housing bubble we saw that investors became overly optimistic. Much of this optimism was fueled by the optimism of other investors and the general public. One study, done by Bikhchandani and Sharma examined herd behavior in financial markets. They illustrate how it can occur with a very clear example. Suppose there are 100 investors that each have their own assessment of investing in a particular market. 20 of the 100 believe it is favorable to invest, whereas 80 believe it is not. Each investor only knows his own individual assessment.
We can see that if these investors shared their information and assessments with each other that they would reach the consensus not to invest. However, these investors do not share information. The result will be that 20 of them will invest. Several of the 80 investors, seeing the behavior of that optimistic 20, may be swayed to also invest in this market. This trend may continue until a majority of the 100 have invested (this is an information cascade).
There are a few key takeaways from this example. First of all, we can see it is the most optimistic investors that are driving the behavior of the herd. If the trend were in the other direction (i.e. to short the market), then it would be the pessimistic investors driving the herd. Since the middle is being swayed by the extremes, this leads to volatility in the market. Lastly, there is no reason to believe that the overly optimistic or overly pessimistic investors have access to better information or are more accurately able to predict the market.
The authors also consider a scenario of three investors who are considering investing in a stock. These investors choose sequentially whether to purchase or reject the stock. Each has his own private information regarding whether the stock is a good investment or a bad investment. What they find is that (similar to the marble example in class) the third investor will copy the behavior of the first two investors if the first two investors make the same choice. This will result in either a buying cascade or a selling cascade. More generally, Bikhchandani and Sharma write, “An individual will be in an invest cascade if and only if the number of predecessors who invest is greater than the number of predecessors who do not invest by two or more.” Once this cascade starts private knowledge becomes irrelevant.
http://www.jstor.org/stable/3867650?seq=1#page_scan_tab_contents