Vernon Smith Hall (formerly Metropolitan Building), #5075
June 14, 2017, 04:00 PM to 01:00 PM
Individual preferences are only partially observable. We only observe choices for those opportunities that are available. However, the design and evaluation of policy many times requires the ability to extrapolate these preferences to new environments. In this dissertation I discuss the issue of extending preference relations from observed behavior in several contexts. The dissertation is organized into three chapters.
The first chapter presents a general representation theorem for the extension of preference relations. That is, it provides the conditions under which the existence of a utility-representable complete extension of an observed incomplete relation can be tested. This allows us to revise the existing revealed preference theory by showing that every revealed preference test can be represented as a set of internal and external consistency conditions.
The second chapter presents criteria under which a set of observed choices can be generated by a complete, transitive, monotone and quasilinear preference relation. I test the empirical content of the quasilinearity of preferences by conducting a laboratory experiment on individual decisions among goods and money. I show that while subjects generally satisfy the generalized axiom of revealed preference, they are no closer to quasi-linear preferences than random choice.
The third chapter analyzes collective choice from a revealed preference perspective. A collective choice function is Pareto rationalizable if there are complete preference relations for each player (satisfying additional desired properties if necessary) such that observed choices are the Pareto efficient outcomes. I characterize the set of Pareto rationalizable single-valued collective choice functions.