| This dissertation contains three separate essays that examine psychological patterns in economic decision-making. The project of merging psychology and economics is motivated by both philosophical (Kahneman et al., 1997) and pragmatic (Chetty, 2015) points of view. Economics as a science ought to be concerned with putting forth the best descriptive theory of choice. By holding consumer choice theories up to actual, idiosyncratic human behavior, we can fine-tune these models to reflect ancillary conditions affecting choice that may not have been immediately apparent to previous theorists, such as asymmetric sensitivity to losses or present-bias. Nevertheless, in broadening our consideration of economic psychology, we should be careful to attend to principles of model simplicity and predictive validity. The most robust economic models are after all often simplest ones that require the fewest assumptions and contextual details, otherwise known as "moving parts". I hope that the research reported in this dissertation properly reflect the guiding principles of both empirical rigour and parsimony.;In Chapters 1 and 2 I examine how people learn about risk and subsequently choose whether to purchase insurance to protect themselves. Chapter 1 reports evidence from a laboratory experiment in which subjects faced real monetary risks and made choices about whether to insure. The primary benefit of studying choices made in a laboratory environment is the ability for the researcher to have complete control over the economic environment and access to high-resolution data on context and beliefs as people make decisions. Subjects in this experiment made insurance choices that reflected a heuristic process of reinforcement learning, which I show to be consistent with anomalies observed in insurance markets for low-probability risks such as flooding.;Chapter 2 more closely examines the role of memory in determining whether people choose to insure, looking specifically at insurance choices made by rural farmers in Gujarat, India. I identify a memory effect by which farmers place disproportionate weight on recent events (floods, droughts or insurance payouts) when deciding whether to insure. Because there is no evidence that more recent events give farmers better information about the current year than non-recent events, there is reason to believe that this memory effect is a byproduct of a behavioral bias, possibly from the representative heuristic.;Chapter 3 is co-authored with Joshua Tasoff. We investigate how overconfidence and production technology interact to influence decision making. This research is important for domains where overconfidence is endemic, such as investing and entrepreneurship. We predict theoretically that productivity-enhancing capital can make an overconfident agent worse off and design an experiment to test this hypothesis. As predicted, we discover that subjects earn less in a skill-based task when provided with certain types of productivity-enhancing capital. |
