UC San Diego

This dissertation examines how people make choices over time and for other people.

In the first chapter, we study when it is possible to link a social planner's preferences across groups of different agents. We propose a preference consistency criterion that relates members of a family of social preferences across domains of different agents; this criterion requires preferences to be identical on domains differing only by adding agents with choice-independent payoffs. We derive additional domain changes for which consistent preferences are invariant, and test adherence to these predictions in an online laboratory experiment. While consistency rates are reasonably high, we document significant differences in consistency across the different types of domain changes. Additionally, we find that participants tend to choose options with higher inequality/lower inefficiency as domain size increases.

In the second chapter, we provide a theoretical approach for investigating attitudes towards intertemporal inequality. We generalize the Pigou-Dalton principle to intertemporal settings by formulating several partial orders on the space of income streams. Three of the partial orders account for payments received over the lifespan of the stream and differ only by how intensely one stream dominates another. A fourth partial order only accounts for the level of inequality experienced in a specific period, rather than over the lifespan of the stream. We then perform a laboratory experiment to distinguish the empirical relevance of these different partial orders and inequality rankings. We find that orders that rank whole streams accurately reflect how participants view streams for themselves, but these views do not translate into how they choose for others. Instead, many of our participants display inequality aversion based on period-wise outcomes.

In the third chapter, we present a laboratory experiment designed to measure both actual and perceived dynamic inconsistency using a novel convex commitment device. We find that participants' demand a great deal of commitment, implying they believe they are significantly dynamically inconsistent, despite evidence of little to no dynamic inconsistency. The results suggest caution when employing commitment devices, as their usage may be unrelated to the problem they are trying to solve.