Essays on Social Influence in Political Economy: How Expectations and Identity Affect Pro-Social Leading and Following

Abstract

By social influence I understand the change in an individual’s thoughts, feelings, attitudes or behaviors that results from interactions with another individual or group. Political, commercial and public health campaigns rely at least partly on influence. Without influence, we have a hard time explaining voter turnout, fads or contagious health behaviors. In my research I focus on pro-social behavior and the de-decentralized provision of public goods, and I ask when and why people are influenced by others as well as when and why people attempt to influence others to “do the right thing”. These questions help us understand human motivation in social contexts, and thus may also help us design policies that can nudge behavior towards more socially desirable, welfare enhancing outcomes. Despite the importance of influence, its study is scattered across disciplines. In my research, I seek to bridge the disciplinary gap through a three-pronged approach. First, I incorporate concepts found in psychology into a decision-theoretic framework. Second, I experimentally test for hypotheses that are derived from this formalization. Third, I use game theory to derive novel conclusions about how aggregate behavior changes when these concepts are incorporated and propose policy recommendations. My dissertation follows parts of this procedure and points to next steps for two psychology concepts: social identity adoption and social expectations. In chapter 1, I write down a unifying model of social identity adoption that integrates different strands in the economics and psychology literature. I provide evidence for the main predictions of this model with a large scale field experiment on charitable giving in Mexico. In chapter 2, joint with Michael Hiscox, we write down a model from which we derive conditions for distinguishing between a social expectations and an altruism explanation to pro-social influence. Results from a laboratory experiment show that most pro-social influence is due to social expectations. In chapter 3, I integrate this social expectations model into a sequential decision setting. I use this to derive a novel model of pluralistic ignorance, and argue that this model explains why uninformed individuals can be leaders in a way past models could not.