Information Cascades among Students vs Financial Professionals
http://www.nber.org/papers/w12767.pdf
Information Cascades among Students vs Financial Professionals
It is not hard to find examples of information cascades and network effects in real life. For instance, when in my high school freshman year, almost all my classmates and I used MSN as an important means of communication with classmates and upperclassmen. However, in the next year, some of us started using Facebook, in addition to MSN, and it didn’t take long until all of the class quit using MSN and use Facebook. The linked article examines if such chain of behavior is manifested differently according to people’s occupations. As the article points out, most of the previous experimental literature on information cascades was limited in terms of the experimental subjects since most experiments were conducted by university researchers and, naturally, used students as the experimental subjects. In contrast, the linked paper compares the behavior of information cascade among students and among people for whom information-based decision-making is an occupational routine: financial professionals.
The experimental design of this research is basically same with the urn experiment covered in the class. The sequential drawing-and-announcing experiment is conducted on professionals from Chicago Board of Trade (CBOT) and on students, as a controls group. One of the findings that I found particularly interesting is that although information cascades are observed in both groups, the CBOT professionals are found to be less affected by the information from others and more focusing on their own personal signals. Also, the professionals are found to be better able to discern the quality of signals associated with other professionals’ announcements and to be less likely to engage in reverse cascades, which end up in “bad” results.
Acknowledging that the behavior of information cascade may appear differently across various types of people allows another dimension of understanding the phenomenon. At least to me, the classroom discussion on information cascades based on Bayesian calculation somewhat gave an expression that information cascade is the outcome of people’s rational decisions. However, although it is true that the phenomenon per se may be observed almost universally, the magnitude and details of the instances may differ according to the settings and individual’s characteristics. Plus, it is a relief that financial professionals are less likely to be affected by information other people’s (announced) decisions since, otherwise, herding behavior which leads to detrimental financial outcomes such as bubbles might have occurred more frequently.