CEU Electronic Theses and Dissertations, 2017
| Author | Mikhailov, Vladimir |
|---|---|
| Title | Essays on Social Networks and Economic Behavior |
| Summary | This thesis consists of 3 unconnected single-authored chapters. Each chapter investigates a particular aspect of social networks' influence on the behavior of economic agents. In Chapter 1 analyze a model in which criminal networks are viewed as embedded in the social network. Agents in the society are assumed to have social preference for or against illegal activity and, accordingly, can help or harm the criminals without actively partaking in crime. I derive predictions for crime participation as a function of centrality in a given network, as well as the effect of network structure on aggregate crime. The equilibrium number of criminals exhibits an inverse-U pattern with respect to public support for crime. If crime is strongly disliked by the society, it is only committed by the most peripheral agents. If, on the other hand, there is social sympathy for the crime, then it is only the most central individuals who become criminals. In terms of network structure, I find that social antipathy towards crime can mean that denser networks exhibit less crime than sparser ones, which reverses the result of Ballester et al. (2006) that denser networks produce more crime on aggregate. I also find that, depending on the society's attitude, an increase in sanctions might fail to deter or even increase aggregate crime. The results reconcile several apparent conflicts between existing models and empirical evidence. Chapter 2 is my job market paper. I use a unique data set which maps out the complete social network within a community of Indian student migrants at a large university in central Kazakhstan to identify endogenous peer effects in assimilation among the community's members. Upon arrival, students are randomly assigned into small academic groups, consisting only of fellow Indians. I use the resulting exogenous variation in social ties to implement a quasi-experimental empirical strategy. Positive peer effects are identified in ability to speak the local language and to acquire local friends. At least a part of the effect is explained by complementarity between assimilation efforts of friends, implying that the peer effects `snowball ' into a social multiplier of 1.4. Finally, assimilation is shown to increase overall GPA, conditional on hours of study. The results suggest that taking advantage of the social multiplier within existing migrant clusters might be a viable alternative to policies, such as settlement quotas, designed to prevent clustering. In chapter 3 I study the problem of a monopolist, who relies on word of mouth in order to diffuse the information about the product through a social network of consumers. The product can be of a certain quality, which is proportional to the probability that the consumer has a positive experience with the good. If the quality is low, the consumer might have a bad experience and choose to give a negative review to friends, discouraging them from purchasing. Discouraged consumers create bottlenecks in the information passage process. I first take quality as exogenously given and show that in highly connected networks, negative WOM makes demand less elastic than the fully-informed case, so the monopolist charges a higher price. Raising the price in this case is a `vaccine' against negative reviews. Later, I endogenize the quality choice and show that if the quality-boosting technology is expensive, then price and quality are substitutes, and the optimal quality goes down with network connectivity, while price goes up. |
| Supervisor | Szeidl, Adam |
| Department | Economics PhD |
| Full text | https://www.etd.ceu.edu/2017/mikhailov_vladimir.pdf |
Visit the CEU Library.
© 2007-2021, Central European University