Essays on Informational Cascades

Author:

Bhalla, Manaswini

Graduate Program:

Economics

Degree:

Doctor of Philosophy

Document Type:

Dissertation

Date of Defense:

June 11, 2009

Committee Members:

James Z Wang, Dissertation Advisor/Co-Advisor
Kalyan Chatterjee, Committee Chair/Co-Chair
Vijay Krishna, Committee Member
Edward James Green, Committee Member
Susan Xu, Committee Member

Keywords:

Informational Cascades
Social Learning

Abstract:

Learning about an underlying uncertainty from observing the actions of other agents is a widespread phenomenon in many social interactions. This thesis looks different aspects of the phenomenon of informational cascades that results due to this social learning. The first essay theoretically analyzes the herd behavior in the presence of payoff externalities. The second essay looks at the optimal product launch and pricing strategy of the monopolist when its consumers learn by observing the unknown quality of the product. The last essay is an empirical investigation of herd behavior among financial forecasters. Chapter 1. Waterfall versus Sprinkler Product Launch Strategy: Influencing the Herd A monopolist must decide the sequence in which a product of unknown quality should be launched across many regions and how it should be priced. The sequencing is important because consumers in regions with later launches can observe the prices and sales in earlier launches. In our model a high price is informative for later consumers as it is an indication that the product is likely to be of high quality. We show that on an average prices decrease over time. Despite the fact that prices decline over time, the profits (per market) actually rise over time. Finally, we identify circumstances in which sequential and simultaneous launches are optimal. We analyze the case where consumers can choose when to consume the product. We find that prices no longer on an average decrease over time. Compared to the case of non strategic consumer case, the monopolist launched the product sequentially for a smaller range of beliefs. Chapter 2. Endogenous Timing, Payoff Externalities and Informational Cascades Previous literature on herd behavior has either looked at models with endogenous timing or payoff externalities but not both. I analyze a more general problem where both issues are analyzed together. To begin with, we theoretically analyze the existence of herd behavior and timing of actions of agents when both informational and payoff externalities are present. In an empirical investigation, we estimate the nature of payoff externalities in the remuneration of financial analysts using I/B/E/S data set. We find that under negative payoff externalities, a herd does not arise and all agents take an action together without any delay. This leads to truthful revelation of all private information. On the other hand, an incentive structure with positive externalities gives rise to a herd led by the most precise agent with a delay in the disclosure of information. From the perspective of firms or investors looking for financial advice, using negative payoff externalities, such as relative performance evaluation, is the most efficient form of incentive structure. Such a structure not only ensures an un-mimicked financial advice but also leads to information disclosure without any delay. Chapter 3. Do Forecasters Herd? Social Learning Among Rational Analysts This paper empirically analyzes two key issues related to herd behavior among financial forecasters. First, we examine if rational analysts learn from each other and hence form an informational cascade. We conduct two tests- a nonparametric and a parametric structural test to infer if analysts learn from the forecasts of previous analysts. We conclude that forecasters do learn from their predecessors when making a prediction. It is also found that most forecasters believe that the predictions of their immediate predecessors contains all information available so far. Next, we test for the nature of payoff externalities in the remuneration system of these forecasters listed with I/B/E/S. We find that the underlying payoff externalities are negative i.e. the benefit from making dissimilar forecasts is higher than that from making similar ones.