Networks

Though our parents implore us to think rationally, it has been established that following the herd mentality occurs even when the trend is contrary to established evidence. For example, the current attitude toward tobacco use is that smoking is generally bad because it increases health risks, and thus the smoking rate remains low. However, Harvard University professor Kip Viscusi has calculated that if the public better appreciated the risks associated with tobacco use (such as lung cancer), the smoking rates would in fact increase.  It is only due to the propagation of misinformation that there is such harsh regulation and prohibition of smoking.  Why is this the case? Economists over the past 20 years have attributed this to cascade theories.

According to UCLA economists Susil Bikhchandani, David Hirshleifer, and Ivo Welch, an informational cascade is defined as a situtation in which “it is optimal for an individual, havin gobserved the actions of others ahead of him, to follow the behavior of the preceding invididual without regard to his own information.”  That is, instead of “buying” information, people will simply follow signals from others in order to conserve time.  There are various examples of cascades in which this occurs.  For example, during the Alar pesticide scare in the late 1980s, public fear of pesticides caused devastating economic loss for apple growers in Washington state who used the pesticide to protect their produce.  There was no scientific basis for the danger of the pesticide in question, but the herd mentality took over and thus the farmers suffered.  While the harsh regulation of cigarettes in the previous example may not be a bad thing, the imposition of bad public policy in the Alar pesticide scare example illustrates how the information cascade effect can have negative consequences.

Pierre Lemieux has proposed a model for this information cascade phenomena.  In this model, an individual observes the decisions of all those before him, and can either adopt or reject a certain behavior based on public decisions and/or his own private signal.  In the beginning of the cascade, individual A will either receive a high signal H or low signal L that corresponds to a signal consistent or contrary with the decision to adopt.  If H, then A will adopt individual B’s outcome, and if L, then A will reject it.  Now individual B will make his choice.  Because he is privy to the public information (in this case, what choice individual A made), B is more likely to choose what A did.  Suppose A got signal H and thus chose to adopt, then B will also choose to adopt.  Consquently, individual C, knowing that both A and B chose H, will also choose to adopt.  Thus the cascade continues.