Network structures and incentives

Resumen

This thesis is divided into 3 distinct chapters: the first, "Outerfactor: a robust approach to the measurement of influence", proposes new measures of the impact of scientific journals, with properties that have no current measures. The second, "Coordination structures", characterizes a set of social networks to ensure the coordination of the society in the actions that most benefit it; and the third, "Sequentiality and simultaneity in bank runs: an experimental approach", shows experimental evidence on how bank runs differ depending on if individuals observe others or, conversely, make their decisions independently. My overall goal in this research is to identify how specific connections modify certain socioeconomic aspects, using the theoretical, the empirical and the experimental methodology. In the first chapter we develop a new framework for analyzing the impact of scientific journals, which we call the research chain model. In this environment we propose a new measure of impact, Outerfactor, which takes into account the indirect impact between journals while assigns to each journal an impact that is independent of its own citation pattern and number of articles. The research chain model also allows us to develop four new measures of mutual impact between journals, extending the concept of citation between journals to the indirect impact environment. Finally, we apply these new measures in a sample of 140 economics journals, showing that Outerfactor provides rankings comparable with other measures while at the same time it has a more robust performance against some possible manipulations. In the second chapter, we develop the concept of observation network, that we use to model the social structure that allows individuals observe the actions they perform. In this context, we characterize the set of observation networks in which there are no coordination failures, networks that we call coordination structures. We show that a necessary and sufficient condition to ensure the inexistence of coordination failures is the existence of a sufficiently large clique, ie a set of individuals in which everyone is able to observe each other. Extending the analysis to situations of private information, we find sufficient conditions for the existence of coordination structures. Moreover, we show other types of coordination structures and social networks that do not guarantee coordination in general, but under certain conditions. Finally, we study the implications of coordination structures for problems of revolts and bank runs. Chapter 3 contains two papers in which we analyze experimentally the case of bank runs, when individuals act sequentially rather than simultaneously. In the first part of the chapter, "On the effects of deposit insurance and observability on bank runs", we study experimentally the effects of both in the emergence of bank runs. We find that the fact of observing actions and the intensity of deposit insurance does not act independently but that the effectiveness of deposit insurance depends on whether people know or not the actions of others when deciding whether keep their savings in the bank. Specifically, we find that the amount of bank runs was lower the higher the deposit insurance when they acted simultaneously. However, additional levels of insurance had no greater effect than an average when individuals observe their actions. In the second part of the chapter, "Do social networks prevent bank runs?", We tested experimentally a model related with the Coordination structures in Chapter 2, for the case of bank runs. In the experiment we found that, although the existence of the observation network had individuals to generate less bank runs, it could also cause them, although such situations were not expected in our theoretical model. We argue that such bank runs are different from those discussed in the literature, based on the fundamental problems of the bank or coordination problems.