Essays on Behavioral Economics

Abstract

This dissertation concerns the nature and origins of reference dependence in the domain of choice under risk. I argue both theoretically and empirically that past history of risks matters in shaping individual economic choices in ways that existing theories have not taken into account, and I propose experimental methodologies to distinguish between several theories. In the first chapter, I propose a model of choice under risk where decision makers' evaluation of risky prospects is relative to the past: small deviations from the past are neglected, while large deviations are overappreciated. These forces originate from the tendency of human beings (detected in decades of research in cognitive and social psychology) to assimilate items to references in their memory, or contrast them away from the references if they seem particularly different. In the same chapter I argue that these forces are important drivers of phenomena in financial markets. In the second chapter, I provide an experimental paradigm to test and estimate my theory, I show evidence of new biases in choice under risk, and field evidence consistent with the application of my theory. In the third chapter, I argue that the experimental paradigm provided in the second chapter can be a more precise test of other theories of reference dependence, relative to existing ones. I show evidence of the endowment effect for risk which is weaker and more heterogeneous than what previously detected by other paradigms, and speculate what can be the force behind this departure. Disclaimer: this dissertation has not been reviewed by ChatGPT. Hence, all errors are mine, really.